DEV Community: Hemapriya Kanagala

Dev Opportunity Radar #18: $50K AI Hackathon, SheBuilds, Verci Fellowship, and Stanford CS336

Hemapriya Kanagala — Fri, 25 Sep 2026 17:48:13 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the Verci Fellowship, Nebius x NVIDIA Global AI Hackathon, and SheBuilds on Lovable, along with Stanford CS336: Language Modeling from Scratch, a course for understanding how language models are built from the ground up.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features two opportunities shared by Dhruv Jani (@dj29). Thank you, Dhruv, for sharing the Snapdragon® AI Lab Build & Present Challenge and the SerpApi India Hackathon 2026. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Verci Fellowship
- 📌 Nebius x NVIDIA Global AI Hackathon
- 📌 SheBuilds on Lovable
📚 Resources Worth Checking Out
- Stanford CS336: Language Modeling from Scratch
🌟 Community Finds
- 📌 Snapdragon AI Lab Build & Present Challenge
- 📌 SerpApi India Hackathon 2026
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Format	Deadline
Verci Fellowship	Verci	Fellowship	Hybrid	October 1
Nebius x NVIDIA Global AI Hackathon	Nebius × NVIDIA	Hackathon	Online	October 30
SheBuilds on Lovable	Lovable	Build Event	Virtual	September 27

Resource Highlight

Resource	Why Check It Out
Stanford CS336: Language Modeling from Scratch	An advanced course covering the process of building language models from the ground up, including Transformers, training, GPU optimization, scaling, data, evaluation, and alignment.

Community Finds

Shared By	Find	Type	Format
Dhruv Jani (@dj29)	Snapdragon AI Lab Build & Present Challenge	AI Challenge	Online
Dhruv Jani (@dj29)	SerpApi India Hackathon 2026	Hackathon	Online & Async

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

As the Radar has grown, there are now quite a few opportunities spread across the different editions. I used to keep all the still-open opportunities here, but I don't think repeating a long list every week makes the editions easier to read.

So I keep this section to the two most recent editions as a quick reminder of opportunities you may have missed. Older opportunities are still available on the Dev Opportunity Radar website, where you can browse the full collection and use the filters to find something specific.

I've already covered the opportunities below in detail, so I won't repeat everything here. If any of them catch your attention, you can go back to the original edition for the full overview, eligibility details, and application links.

Opportunity	Organization	Type	Format	Deadline	Featured In
Code in Place X	Stanford University	Programming Course	Virtual	September 28	Edition #17
Build, Ship, Shape: Amazon Developer Hackathon	Amazon	Hackathon	Online	October 24	Edition #17
Dev3Pack Hackathon	Dev3Pack	Hackathon	Hybrid	October 30	Edition #17
AI for Education Fellowship	Worldwide Studios	Fellowship	Virtual (Global)	Rolling	Edition #16
AWS Student Rewards	AWS Builder Center	Student Rewards Program	Online (Global)	Open	Edition #16

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Verci Fellowship

Who it's for: Young builders, students, recent graduates, and people building projects or communities.

What it's about: The Verci Fellowship is a one-semester program bringing together 20 people who are building startups, projects, and communities.

The fellowship focuses heavily on community and mentorship. Each fellow is matched with a mentor, joins a small weekly mastermind group, and gets access to office hours and workshops with founders, operators, and investors.

There is also a Catskills retreat in November where the cohort spends a weekend together, along with support for fellows who want to organize things in their own communities such as hackathons, dinners, or demo nights.

Another part I found interesting is the focus on building in public. The program includes workshops and accountability groups to help fellows share what they are working on and actually ship something during the semester.

The fellowship is open to people who are currently in college, on leave, dropped out, or recently graduated, with the program looking for people roughly under 24 who have a strong interest in technology, are creative and curious, and are already building things on campus or in their communities.

Applications are rolling, but the current application form says applications close on October 1, 2026 at midnight PST.

Format: Hybrid / Online + New York and LA opportunities

Cohort: 20 fellows

Program: October – December 2026

Application Deadline: October 1, 2026 at midnight PST

Eligibility: Currently in college, on leave, dropped out, or recently graduated; roughly under 24

What you get: Individual mentorship, weekly mastermind groups, workshops and office hours, Catskills retreat, community/event support, cohort Slack, building-in-public support, and optional opportunity matching

Application: Name and background + social profiles + semester schedule + work/portfolio + short answers + 1–2 minute personal video

🔗 Apply to the Verci Fellowship

📌 Nebius x NVIDIA Global AI Hackathon

Who it's for: AI builders, ML engineers, developers, founders, and anyone interested in building practical AI systems.

What it's about: The Nebius x NVIDIA Global AI Hackathon is a global online hackathon where participants build an AI project using NVIDIA open source models on Nebius Token Factory or Nebius AI Cloud.

There are four main tracks, so you can choose the type of project that fits what you want to build:

Coding and Agentic Engineering: Build coding agents and developer tools.
Best Apps and Agents: Build an application or agent designed to solve a real problem.
Personal AI: Build a private, always-on AI assistant with memory, skills, and access to your chosen tools and information.
Physical AI: Build robotics, IoT, or other systems that can sense and act in the physical world.

The main requirement is that your submission runs on Nebius Token Factory or Nebius AI Cloud and uses at least one NVIDIA open source model.

I wanted to include this one because there is quite a bit of room in what you can build. You can work on an AI agent, developer tool, personal assistant, application, robotics project, or something else that fits one of the tracks.

There are also Builders & Brews events in different cities where participants can join technical onboarding, workshops, team formation, build sessions, and get feedback while working on their projects.

For the submission, you'll need a working project, public code repository, project description, working demo or test build, and a public demo video of three minutes or less. Your repository also needs an open source license and setup instructions.

There is $50,000+ in prizes, including a $20,000 grand prize, $10,000 for second place, $6,000 for third place, track prizes, city awards, and additional prizes.

Format: Online | Global

Deadline: October 30, 2026 at 10:30 PM IST

Tracks: Coding & Agentic Engineering | Best Apps & Agents | Personal AI | Physical AI

Requirement: Use Nebius Token Factory or Nebius AI Cloud + at least one NVIDIA open source model

Submission: Working project + public repository + demo + ≤3-minute demo video

Prize Pool: $50,000+ in prizes

🔗 Join the Hackathon

📌 SheBuilds on Lovable

Who it's for: Women 18+ who have an idea they want to turn into a working product, including students, first-time makers, founders, and experienced developers.

What it's about: SheBuilds on Lovable is a global community of women building with AI through virtual build events, webinars, and community spaces.

Season 04 is a 72-hour virtual build event where participants use Lovable to turn an idea into a working product alongside mentors, community champions, and a few hundred other builders.

You don't need to know how to code to apply. You can come with an idea, sketches, research, or a plan, and the actual building starts when the event kicks off.

You can participate solo or with one co-builder. Around 400 participants are expected to be selected for this season.

There are also live prep sessions from October 6–14 covering things like figuring out what to build, making your product stand out, and getting it in front of people. Accepted participants also get Lovable credits to use during the build.

The build starts on October 15 and submissions close on October 17. The organizers will announce the prizes at kickoff, so the prize details for Season 04 have not been announced yet.

Format: Online | 72-hour build event

Build Dates: October 15–17, 2026

Applications Close: September 27, 2026 at 11:59 PM PT

Decisions: October 5, 2026

Eligibility: Women 18+ with an idea they want to build

Team Size: Solo or 1 co-builder

Cost: Free | Coding Required: No

Expected Cohort: Around 400 participants

Prep Sessions: October 6–14

Prizes: To be announced at kickoff

🔗 Apply for Season 04

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Stanford CS336: Language Modeling from Scratch

If you want to go beyond using LLMs and actually understand how language models are built, Stanford's CS336: Language Modeling from Scratch is a really interesting resource to explore.

The course walks through the process of building language models from the ground up, including data collection and cleaning, tokenization, Transformer architecture, training, GPU optimization, scaling, evaluation, inference, alignment, and reasoning.

What makes this course different is how implementation-heavy it is. The assignments have students actually build components of a Transformer language model, train a small language model, optimize attention with Triton, work with distributed training, process pretraining data, and apply post-training methods such as supervised fine-tuning and reinforcement learning.

The course provides lecture recordings, assignments, lecture materials, and code, so you can follow along even if you're not enrolled at Stanford.

One important thing to know before jumping in: this is an advanced course. The course expects proficiency in Python, experience with deep learning and systems optimization, familiarity with PyTorch, and a foundation in calculus, linear algebra, probability, statistics, and machine learning.

You will also be writing a lot of code, so this is much more suitable if you already have a solid technical foundation and want to understand the engineering behind language models.

Format: Self-paced online resource

Focus: Language models, Transformers, training, systems optimization, scaling, data, evaluation, alignment, and reasoning

Level: Advanced

Prerequisites: Python + PyTorch/deep learning + systems optimization + calculus + linear algebra + probability/statistics + machine learning

Includes: Lectures + recordings + assignments + code + course materials

🔗 Explore the CS336 Course Website

🎥 Watch the Lecture Recordings

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, Dhruv Jani (@dj29) sent me two opportunities that I thought were worth including:

📌 Snapdragon AI Lab Build & Present Challenge

Shared by Dhruv Jani

The Snapdragon® AI Lab Build & Present Challenge by Qualcomm is an individual online challenge for participants in India to build or propose an AI solution designed, developed, or optimized for Snapdragon-powered HP PCs.

You can submit a new project or significantly modify an existing project by adding AI models from Qualcomm AI Hub or other open-source platforms.

The challenge is open to Indian residents aged 18+ who own a Snapdragon-powered laptop.

Format: Online | Individual

Deadline: September 30, 2026 at 11:59 PM IST

Eligibility: India residents, 18+

Requirement: Snapdragon-powered laptop

Rewards: HP laptops, Qualcomm internship opportunity for the first award subject to eligibility, executive mentorship, and showcase opportunities

🔗 Learn More & Apply

📌 SerpApi India Hackathon 2026

Shared by Dhruv Jani

The SerpApi India Hackathon 2026 is an online and asynchronous hackathon where participants can build an app, AI agent, or open-source integration using live search data from SerpApi.

You can participate solo or in a team of up to five, use any programming language or framework, and work from anywhere in India. Existing projects are also allowed if the submitted version makes meaningful use of SerpApi.

There are six project tracks, and AI-assisted development is allowed.

The total prize value is ₹3 lakh+, including cash awards, SerpApi credits, track awards, community awards, and credits for every valid submission.

Format: Online & asynchronous

Deadline: October 10, 2026 at 11:59 PM IST

Eligibility: India residents, 18+

Team Size: 1–5 members

Requirement: Meaningful use of SerpApi search data

Prize Value: ₹3 lakh+

Submission: Public GitHub repository + project description + demo video under 3 minutes

🔗 Submit Your Project

Thank you to Dhruv Jani (@dj29) for thinking of the Radar and sharing this one with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

If you discovered an opportunity through the radar, applied to something, joined a community, attended an event, or simply found a resource you hadn't seen before, I'd genuinely love to hear about it.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

If you'd like to share an update, feel free to leave a comment on this edition. With your permission, I may feature it in a future 💙 Reader Updates section and tag you so other readers can celebrate your journey too.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

Every week, I spend time searching for opportunities, resources, and communities that I think deserve a little more attention. But one of my favorite parts of this series isn't the research. It's seeing what happens after an edition is published.

Seeing someone discover an opportunity, apply to a program, share a resource, suggest a Community Find, or come back to tell us what they learned reminds me why I started Dev Opportunity Radar in the first place.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Whether you've been reading since the very first edition or this is your first time here, thank you for being part of the journey. Every comment, suggestion, and shared opportunity helps make this series better than I could build on my own.

If you ever come across an opportunity, resource, event, community, or anything else you think deserves more attention, I'd love for you to share it in the comments. And if Dev Opportunity Radar helped you discover something exciting, I'd love to hear that too.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

If this is your first time discovering the series, you can explore every edition, browse opportunities by category, discover Community Finds, and catch up on Reader Updates on the Dev Opportunity Radar website.

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

What DEV Series Are You Reading or Writing? Let’s Make Them Easier to Find

Hemapriya Kanagala — Wed, 23 Sep 2026 18:37:32 +0000

TL;DR

DEV has so many interesting series, but they can be easy to miss when we don't follow the author or happen to come across a post at the right time.

Until we have a dedicated way to discover series, I thought we could use this thread to share the series we are writing or reading so more people can find them.

If you're writing a series: Share the title, a short introduction, its status (ongoing or completed), and the link.

If you're reading a series: Share the title, what you find interesting about it, and the link.

If you're planning a series but haven't started yet, bookmark this thread and come back when you publish the first part.

You can share as many series as you want. There are no limits here 😄

A Small Problem I Keep Noticing
So, Let’s Share Them Here
Until DEV Has a Dedicated Series Section

A Small Problem I Keep Noticing

There are many great series on DEV. Some are educational, some document a learning journey, some explore a specific technology, and others are simply fun to follow.

But finding these series consistently can be difficult. Unless we follow the author, discover one of their posts, or happen to see it in our feed, we might never know that the series exists.

And by the time we discover it, several parts may already be published. I know I have probably missed some really interesting series simply because I didn't know they were there.

So, Let’s Share Them Here

If you're writing a series, feel free to share it in the comments. You can include the series title, a short introduction about what it covers, whether it is ongoing or completed, and a link to the series or its first article.

If you're reading a series that you think others might enjoy, share that too. Tell us what makes it interesting or what you've learned from it.

If you're planning to start a series, you can bookmark this thread and come back once you've published the first part. That way, people can discover it early and follow along as you continue writing.

You can share as many series as you want. There are no limits here 😄

Until DEV Has a Dedicated Series Section

I remember @dj29 raising an issue in the Forem repo about a possible series discovery feature, so fingers crossed that we get something like that in the future 🤞

I know DEV already has so many great features, and I also understand that building something new takes time and effort. This is not meant as a complaint at all. I'm always grateful to the DEV team for creating and maintaining a space where we can learn, write, and connect with each other.

Until then, maybe we can use this thread as our own little discovery space. There are probably many series worth reading that we simply haven't found yet.

What series are you writing or reading right now? Share them below so we can discover something new together.

Dev Opportunity Radar #17: $138K Amazon Hackathon, Stanford's Code in Place X, and Dev3Pack Hackathon

Hemapriya Kanagala — Fri, 18 Sep 2026 16:44:35 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features Code in Place X, the Build, Ship, Shape: Amazon Developer Hackathon, and the Dev3Pack Hackathon, along with Coding Interview University, a free study plan for software engineering and technical interview preparation.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features an opportunity shared by Dhruv Jani (@dj29). Thank you, Dhruv, for sharing the Dogfood | 72-Hour Hackathon. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Code in Place X
- 📌 Build, Ship, Shape: Amazon Developer Hackathon
- 📌 Dev3Pack Hackathon
📚 Resources Worth Checking Out
- Coding Interview University
🌟 Community Finds
- Dogfood: 72-Hour Hackathon
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Format	Deadline
Code in Place X	Stanford University	Programming Course	Virtual	September 28
Build, Ship, Shape: Amazon Developer Hackathon	Amazon	Hackathon	Online	October 24
Dev3Pack Hackathon	Dev3Pack	Hackathon	Hybrid	October 30

Resource Highlight

Resource	Why Check It Out
Coding Interview University	A free, self-paced study plan covering data structures, algorithms, computer science topics, coding practice, and software engineering interview preparation.

Community Finds

Shared By	Find	Type	Format
Dhruv Jani (@dj29)	Dogfood	72-Hour Hackathon	Software Development Hackathon

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

Opportunity	Organization	Type	Format	Deadline	Featured In
AI for Education Fellowship	Worldwide Studios	Fellowship	Virtual (Global)	Rolling	Edition #16
AWS Student Rewards	AWS Builder Center	Student Rewards Program	Online (Global)	Open	Edition #16
Bharat Builds Tour	WeMakeDevs × AWS Builder Center	Student Hackathon Series	Online (India) / In Person	September 17–20	Edition #15

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Code in Place X

Who it's for: Anyone 18+ who wants to learn programming, or people who already know Python and want to volunteer as a teacher.

What it's about: Code in Place X is a new free course from Stanford that teaches programming while also helping students practice explaining what they build.

The course is six weeks long and you'll meet every week on Zoom with a small group of students and a volunteer teacher who will guide you through the material. You don't need any previous programming experience, and the course is completely free.

I wanted to include this one because I like that it is not only about writing code. Students also get to practice explaining their thinking and what they built, which is a useful skill whether you're learning programming for the first time or already working with AI coding tools.

The course also has an interesting approach to AI. Students can use AI as part of their learning, but they are expected to come to their weekly section ready to show what they learned and share how they are using AI. So the course is also exploring what it means to learn programming when AI can already write a lot of code for us.

There is also a volunteer teacher option. If you already know how to write a simple Python program and can explain your thinking, you can apply to lead a small group. You don't need to be a professional engineer or have teaching experience. Teachers spend around 4 hours per week and receive lesson plans and section problems to work with.

The course itself takes around 7 hours per week for students and begins on October 12, 2026.

Format: Online | Duration: 6 weeks

Class Begins: October 12, 2026

Orientation: October 7, 2026

Time Commitment: Around 7 hours/week for students | Around 4 hours/week for volunteer teachers

Cost: Free

Eligibility: 18+ | No prior programming experience required for students

Applications Close: September 28, 2026

Language: English

🔗 Learn More & Apply

📌 Build, Ship, Shape: Amazon Developer Hackathon

Who it's for: Developers, students, AI builders, and anyone interested in building applications for Amazon devices and services.

What it's about: The Build, Ship, Shape: Amazon Developer Hackathon brings Fire TV, Alexa+, Bee, and Ring together in one global hackathon. You can choose one of the four tracks and build a working project using the developer tools, APIs, SDKs, or simulators for that track.

I wanted to include this one because there are quite a few different ways to approach the challenge. You can build for Fire TV, Alexa+, Bee, or Ring, depending on what you're interested in. There are also additional AWS Builder and Open Source mini challenges that you can enter alongside your main track.

For example, the Fire TV track lets you build an app using React Native, web technologies, Android, or another framework that runs on Fire OS or Vega OS. The Alexa+ track focuses on building an MCP server or Agent Skill, while Bee projects use data from a Bee device or Apple Watch running Bee software. Ring projects can use Ring APIs, SDKs, simulators, or devices.

You also don't necessarily need to already have experience with Amazon's developer tools. The hackathon provides documentation, sample code, simulators, and live office hours to help participants get started.

The submission needs to include a working project, GitHub repository, and a demo video under three minutes, along with product feedback on the Amazon tools you used.

There is a $138,000 cash prize pool, with prizes across the four main tracks and the AWS Builder and Open Source mini challenges. Track winners also receive a 1:1 meeting with the Amazon Developer team, depending on the track.

One thing to check before choosing your project is the specific requirements for each track. Some tracks require you to actually use the relevant technology in your code or demo, so simply mentioning a particular Amazon service isn't enough.

Format: Online | Global

Deadline: October 24, 2026 at 12:30 AM IST

Team Size: Not specified in the information provided

Main Tracks: Fire TV, Alexa+, Bee, Ring

Mini Challenges: AWS Builder, Open Source

Prize Pool: $138,000 in cash + additional AWS credits and other prizes

Submission: Working demo + GitHub repository + under-3-minute demo video + product feedback

Eligibility: Above the legal age of majority in your country of residence | Global, subject to standard country/territory exclusions

🔗 Learn More & Join the Hackathon

📌 Dev3Pack Hackathon

Who it's for: University students and student builders interested in building on Solana. You can participate solo or in a team of up to four.

What it's about: Dev3Pack Hackathon is a three-day hybrid hackathon where students from around the world can build on Solana, either online or at one of the local build stations.

The hackathon runs from October 30 to November 1, 2026, but there is also a free Solana Rust Bootcamp running from October 6 to 30. So if you're new to Rust or Solana, you can use the weeks before the hackathon to learn and get ready.

I wanted to include this one because it has a little more of a community and learning side than just the three-day hackathon. You can join a local build station if there is one near you, meet other students and mentors, or participate completely online if you prefer.

There are currently 40+ build stations listed around the world, including Mumbai, along with locations across Asia, Africa, Europe, and the Americas.

The project needs to be open source and must use Solana if you want to be eligible for prizes and the accelerator. You can also use AI coding tools while building, as long as the project is yours and you understand and can explain what you built.

After the hackathon, top teams can be invited to the Bridge Accelerator, which runs from November through January. The exact prize tracks and amounts for this edition are still to be announced.

There are also some builder perks available to participants, including Noah AI credits, v0 credits, and Cursor Pro for select countries, along with other perks listed in the participant dashboard.

Format: Hybrid | Online or local build station

Dates: October 30 – November 1, 2026

Preparation: Solana Rust Bootcamp | October 6–30

Duration: 3 days

Team Size: Solo or up to 4 members

Location: Worldwide | 40+ build stations

Cost: Free

Project Requirement: Open source + Solana

AI Tools: Allowed

Prizes: Tracks and prize amounts to be announced

After the Hackathon: Top teams can be invited to the Bridge Accelerator

🔗 Learn More & Apply

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Coding Interview University

If you're preparing for software engineering interviews and don't know where to start, Coding Interview University is a free study plan that puts a lot of the topics you need to learn in one place.

It covers things like data structures, algorithms, Big-O, trees, graphs, sorting, dynamic programming, networking, and coding interview practice. It also has sections on resumes, finding a job, interview preparation, and other parts of the interview process.

What I like about this resource is that it gives you a path to follow instead of just giving you a list of random topics. You can go through the topics step by step and practice coding problems along the way.

There are also a lot of free resources linked throughout the plan, including videos, books, courses, and coding practice sites.

It is a pretty big study plan though, so you definitely don't need to finish everything. You can pick the parts that make sense for where you are in your interview preparation.

Format: Free, self-paced study plan

Covers: Data structures, algorithms, coding problems, computer science topics, and interview preparation

🔗 Explore Coding Interview University

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

Dogfood: 72-Hour Hackathon

Shared by Dhruv Jani (@dj29).

Dhruv shared the Dogfood | 72-Hour Hackathon by Hackathon Raptors, and I thought this was an interesting one to include because the challenge itself is quite different from a usual hackathon.

Instead of letting everyone build different projects, everyone is working on the same product: an open-source, self-hostable platform for running hackathons, including registration, team formation, submissions, judging, scoring, and results.

What I found interesting is that the goal isn't just to build a working demo. The winning project will be forked, self-hosted, and used for future Hackathon Raptors events. Teams also keep ownership of their work.

The technical requirements are quite specific too. The platform needs to run locally with docker compose up and work without a hosted database, external API, cloud account, or network connection. The challenge also covers things like role isolation, judging, score normalization, voting, APIs, testing, and documentation.

There are four tiers to work through, but you don't have to try to build everything. The organizers mention that a clean and correct lower tier is better than a broken higher tier, which I thought was a good approach for a 72-hour challenge.

Who it's for: Developers, students, engineers, open-source contributors, and anyone interested in software development, backend systems, APIs, DevOps, security, or data and algorithms.

Format: Online

Duration: 72 hours

Event Date: September 25–28, 2026

Team Size: 1 to 4 participants

Prize Pool: ₹2,50,000

🔗 Learn More & Register

Thank you to Dhruv Jani (@dj29) for thinking of the Radar and sharing this one with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

What Happens When AI Outgrows the Tests We Use to Measure It?

Hemapriya Kanagala — Mon, 14 Sep 2026 17:12:10 +0000

TL;DR

GPT-6 Astra has started another familiar AI conversation. The model is more capable, Jensen Huang said on X that "AGI has arrived," and social feeds quickly moved between excitement, curiosity, and fear about what comes next.

I understand why the reactions are strong. AI is changing quickly, and some of the things these models can do now would have sounded surprising not long ago. But I have also noticed how quickly "AI is becoming much more capable" can turn into "developers will not be needed" or "CSE is finished." That feels like a much bigger conclusion.

For me, Astra raises another question that is easier to miss:

How do we actually know an AI model is getting better?

We usually look at benchmarks, and they are useful. But benchmarks can age too. Some become saturated, some use proxies for capabilities that are difficult to measure directly, ground truth can be complicated, and evaluation methods can change. A benchmark score also does not automatically tell us how useful a model will be for our own work.

Astra's system card gives several examples of this, including older evaluations becoming saturated or being considered for retirement, newer evaluations using more realistic or experimental data, and evaluation methods changing over time.

That matters for developers too. If AI makes writing code cheaper and faster, software engineering does not suddenly disappear. Requirements, architecture, system design, security, verification, tradeoffs, maintenance, and deciding what should actually be built still matter.

AI will change the work. I think that part is pretty clear.

The more useful question is what we do with that change, and whether we are measuring AI well enough to understand what is actually happening.

Another AI launch, and the conversation gets loud very quickly
A benchmark gives us a number. What does the number really tell us?
Every benchmark has a lifespan
92% according to what?
The number can change because the test changed
One answer is starting to tell us less
So what should we look at when we see a benchmark?
What does all of this mean for developers?
We can take AI seriously without making every headline the final answer
The models are changing. Our way of looking at them has to change too.
I Would Love to Hear From You
References
🤝 Let's Stay Connected

Another AI launch, and the conversation gets loud very quickly

Whenever a major AI model arrives, there seems to be a familiar sequence. The model launches, people try it, benchmark results appear, comparisons start, and then the bigger questions arrive.

With Astra, NVIDIA CEO Jensen Huang also said on X that "AGI has arrived" while congratulating OpenAI on the launch.

As you would expect, people started interpreting that statement in very different ways. Some saw it as a major milestone, some questioned whether Astra actually meets their definition of AGI, and others immediately started thinking about what this could mean for jobs and software development.

From what I have been seeing in my own feed, there is also a lot of fear around where developers fit into all of this.

I am not saying fear is wrong. There are real risks with increasingly capable AI, and I think those risks deserve serious attention.

But I think there is a big difference between saying "this technology is changing quickly" and saying "developers are no longer needed" or "there is no need for CSE anymore."

That feels like a much bigger conclusion to draw from a capability improvement.

Software development will change. There is very little doubt about that. But when producing code becomes easier, other parts of the work can become more important too.

Understanding the problem, designing the system, making architecture decisions, working through requirements, thinking about security and reliability, verifying what was produced, understanding tradeoffs, and knowing when an AI-generated solution is simply the wrong solution are all still part of building software.

None of those become less important just because writing the code becomes faster.

So while everyone is asking how capable the new model is, I started wondering about something slightly different:

How are we actually measuring that capability?

A benchmark gives us a number. What does the number really tell us?

There is something very convenient about benchmarks. You see Model A at 82% and Model B at 88%, and it suddenly feels like we have a clear answer about which one is better.

And honestly, benchmarks are useful. Without them, comparing AI systems would be much harder. Having a common test gives us a starting point, and that is valuable.

The part that is easy to forget is that a benchmark is a measurement tool. It gives us evidence about a model's capability, but it is not the capability itself.

Think about a thermometer. It tells you the temperature, but the thermometer is not the temperature. In the same way, a benchmark can tell us something about what a model can do, but it cannot capture everything the model is capable of doing in the real world.

And this becomes more important as models improve, because the test itself can eventually become less informative.

A benchmark that once showed a clear difference between models might eventually have almost every strong model scoring near the top. At that point, the number is still a number, but it may not tell us as much as it used to.

Every benchmark has a lifespan

One of the things I found interesting in Astra's system card is how openly it talks about evaluations changing over time. Some earlier evaluations have become relatively saturated, some are being considered for retirement, and newer evaluations are being introduced because the older ones may no longer give enough signal.

That actually makes a lot of sense. Imagine a test where the strongest models are now scoring 97%, 98%, and 98%. Those are obviously impressive results, but if most capable models are already close to the ceiling, how much are we really learning from that one percentage point?

The benchmark may not have become bad. It may have simply reached the point where it is no longer good at showing the differences we care about.

In a way, that means the benchmark did its job. It helped us see progress when the models were still further away from the ceiling, and now the models have moved far enough that we need a different way to separate them.

So when I see an evaluation being retired or replaced, I do not necessarily see that as a failure of the benchmark. Sometimes it is actually a sign that the technology has moved forward.

The models moved, so the test needs to move too.

And I think that idea is going to matter more and more as AI systems keep getting better.

92% according to what?

This is probably my favorite question to ask when I see a benchmark score.

92% according to what?

For some tasks, the answer is easy. A math problem has a known answer. A programming problem can be checked with tests. A factual question can often be compared against a trusted source.

But things get more complicated when we are evaluating scientific reasoning, biological predictions, or longer tasks where there may not be one obvious correct answer.

That is where ground truth comes in. In simple terms, ground truth is the answer or evidence we treat as the correct reference when we evaluate a model. Depending on the problem, it might come from experts, experimental data, a simulation, or some other trusted source.

And sometimes we cannot directly measure the capability we care about, so we use a proxy. A proxy is something we can measure that is expected to tell us something about a harder-to-measure capability.

Astra's system card gives an example from biology. One of its evaluations looks at AAV capsid packaging prediction. AAV stands for adeno-associated virus, and a capsid is the protein shell around a virus. In this evaluation, the model is asked to predict how well different capsids may perform in packaging. The result is being used as a proxy for certain broader biological design capabilities.

Using a proxy is not automatically a problem. In many areas, it is simply a practical way to measure something that would otherwise be very difficult or expensive to test directly.

But there is an important difference between saying "the model got better at this prediction" and saying "the model got better at the real-world capability this prediction represents."

Those two things can be connected without being exactly the same.

The system card also discusses a related issue with BPNet, a model used to make biological predictions. If the predictions from the original system contain errors or assumptions, using those predictions as the test's reference can create a strange situation where a future AI system might become better at reproducing those patterns without necessarily becoming better at the underlying biological task we actually care about.

That is a subtle measurement problem, but it is an important one.

The score can be real, the improvement can be real, and we can still have questions about what that improvement actually represents.

This is also why Astra's newer biological evaluations using unpublished experimental data and wet-lab-validated evidence are interesting. Wet-lab validation basically means the results have been checked through actual laboratory experiments rather than relying only on a computational prediction. That gives the evaluation a closer connection to the real-world behavior it is trying to measure.

So when I see a benchmark saying 92%, I am still interested in the number.

I just want to know what is behind it.

92% according to what?

The number can change because the test changed

There is another part that can easily get lost when we look at benchmark charts: the test itself can change.

A benchmark score does not appear on its own. There is a dataset behind it, instructions for the model, a way of grading the answer, a scoring method, and sometimes specific tools or model settings. If any of those things change, the meaning of the final number can change too.

Astra's system card points this out directly. Policies, graders, datasets, evaluations, and other measurement details can evolve over time, which means older results should generally not be treated as directly comparable to newer ones without looking at what changed.

So when we see something like:

Model A: 90%
Model B: 95%

the five-point difference might be meaningful. But before using it to make a big claim about one model being much better than another, it is worth checking whether both numbers came from the same kind of evaluation setup.

There is even a smaller detail that can affect some results: the length of the answer.

In some open-ended evaluations, a longer answer can have more opportunities to satisfy the different things a grader is looking for. That means a model could potentially score better simply because it said more, even when a shorter answer might have been more useful to a person. Astra's system card discusses this issue and reports length-adjusted scores for some evaluations to account for it.

It may sound like a small technical detail, but these are exactly the kinds of details that matter when we are trying to understand what a benchmark score is really telling us.

The number is important.

But the way we arrived at the number matters too.

One answer is starting to tell us less

There is also a change happening in how we actually use AI. We are not only asking a question and waiting for an answer anymore. AI systems can work through multi-step tasks, use tools, interact with software, deal with incomplete instructions, and adjust when the requirements change.

Astra's launch material talks about this kind of behavior too. It describes the model handling ambiguity better, asking focused questions when an answer could change the outcome, and continuing with other work while waiting for a response. When the decision is more consequential, it can wait for the user's input instead of simply guessing.

That kind of behavior is difficult to capture with one isolated question.

If I am asking an AI system to help with a 30-minute workflow, I probably care about much more than whether it got one answer right. Did it understand what I was actually trying to do? Did it make reasonable decisions along the way? Did it handle a requirement that changed halfway through? If something went wrong, could it recover? And at the end, did the result actually work?

The Astra system card also includes more realistic workplace-style evaluations involving computer and browser tasks, ambiguous instructions, and complex permissions.

That feels closer to how many of us experience AI now.

A model can be excellent at answering individual questions and still struggle when it has to carry a task from beginning to end. The opposite can also happen. A system may not look extraordinary on one isolated benchmark but perform surprisingly well when given the tools and context it needs.

That is why I think the evaluation question is slowly changing too.

Instead of only asking "Did the model get the answer right?", we may also need to ask "How well did it handle the whole task?"

And that is a much harder thing to measure.

So what should we look at when we see a benchmark?

I do not think we need to become experts in evaluation methodology every time a new model launches. But there are a few questions that can make a benchmark score much easier to understand.

1. What is actually being measured?

Is the benchmark measuring the capability we care about, or is it using a proxy, something that is expected to tell us about that capability?

2. Is the benchmark still difficult?

If most strong models are already close to the top score, the benchmark may have become saturated, meaning it is no longer giving us much room to see meaningful differences between models.

3. Where does the ground truth come from?

What are we treating as the correct answer? Is it a fixed answer, expert judgment, another model, a simulation, or real-world evidence?

4. Did the evaluation change?

Did the dataset, grader, scoring method, prompts, tools, or model setup change? Even small changes can affect how we should interpret the final number.

5. Does the result transfer to the work I actually care about?

This one is especially useful if you are a developer. A model can perform extremely well on a general coding benchmark and still not be the best model for your actual codebase, team, or workflow.

For that, your own small evaluation can be surprisingly useful.

Take 20 or 30 real tasks from your workflow and try the models on them. Then look at the things that actually matter to you. Did the code work? Did the model understand the requirements? How did it handle changes? How much did you have to correct? Could it recover when something went wrong? How long did it take, and what did it cost?

You do not need to throw away the big benchmark scores. They can still give you a useful general picture.

Your own tasks simply add another piece of evidence, and sometimes that is the piece that matters most.

What does all of this mean for developers?

This is where the benchmark discussion connects back to something I have been seeing a lot around Astra: "AI will replace developers."

I can understand why people are asking that question. If a model can write code, understand a codebase, use tools, work through a task, and handle changes along the way, it is reasonable to wonder what software development will look like a few years from now.

I do expect the work to change. Some things will become much faster, some tasks may need less human effort, and there will probably be skills that become less important as the tools improve. At the same time, other skills can become more important.

If producing code becomes much easier, for example, the difficult part may move toward understanding what should actually be built and how it should be built.

What are the real requirements? What constraints do we have? How should the system be designed? Will the architecture hold when the system grows? Is the solution secure and reliable? How do we verify what the AI produced? What happens when the requirements change six months from now? And does the whole thing actually solve the user's problem?

These questions do not disappear because an AI can produce hundreds of lines of code quickly.

In some cases, they may become even more important because producing several possible implementations becomes easier. Someone still has to understand the options and decide which one makes sense for the situation.

That is why the idea that "CSE is no longer needed" feels too extreme to me. Computer science and software engineering have always been about more than writing code. Code is one part of building software, not the entire process.

There will definitely be changes, and some of those changes may be significant. But I would rather watch how the work actually evolves than decide too early that the whole field is going away.

Maybe the more useful question for developers is not "Will AI replace me?", but "As AI takes over more of the work, what should I become better at?"

We can take AI seriously without making every headline the final answer

I do not want to dismiss the concerns around AI. There are genuine questions around safety, misuse, security, reliability, employment, and how much autonomy we should give these systems. Those are not small questions, and they deserve serious attention as the technology becomes more capable.

At the same time, I do not think every capability improvement needs to turn into a prediction about the end of a profession. When I see that AI can write code, I see a technology that is changing software development. I do not think the only possible conclusion is that developers are finished.

The same goes for increasingly complex AI systems. Instead of jumping straight to whether humans will be needed, I find it more useful to ask which parts of the work we can trust AI with, where we still need people involved, and what kind of oversight makes sense.

And when I see a benchmark score go up, I want to understand what is behind that number too. What was tested? How was it tested? Is the benchmark still giving us a useful signal? Does that improvement show up in the kind of work people actually do?

For me, that makes the conversation more interesting. We can be excited about what these models are doing and still question the claims around them. We can take the risks seriously without assuming the most extreme outcome, and we can recognize real progress without treating every new headline as the final answer.

There is probably a lot we still do not know about where all of this is going.

And that is okay. We are watching it change in real time.

The models are changing. Our way of looking at them has to change too.

That is probably what I keep coming back to with Astra. The model is becoming more capable, but at the same time, measuring that capability is getting more complicated. Benchmarks can saturate, proxies have limits, ground truth is not always simple, and evaluation methods have to keep evolving too.

Maybe our questions need to evolve along with them.

Instead of only asking "Will AI replace developers?", we can ask "Which parts of development are changing, and what skills are becoming more valuable?"

Instead of "Is CSE dead?", we can ask "What does computer science look like when producing code becomes much easier?"

And instead of only asking "Is AGI here?", we can ask "What capabilities are we actually talking about, how are we measuring them, and what evidence do we have?"

We are going to keep seeing stronger models, impressive demos, and very strong opinions around them. Some predictions will probably turn out to be right, some will go too far, and some things will surprise all of us.

That is part of watching a technology change this quickly.

For me, the useful thing is to stay curious, look at the evidence, take the risks seriously, and also pay attention to what is actually changing in the work we do.

Because after all the benchmarks, headlines, and predictions, maybe the question worth asking is simply:

"What changed, how do we know, and what does that mean for the work we care about?"

And then:

"Okay. This is what the technology can do now. What should we build with it?"

I Would Love to Hear From You

We are all watching the same AI developments, but we do not always see them in the same way.

Maybe you think the current progress is being underestimated. Maybe you think some of the excitement or fear is going too far. Or maybe you are already seeing changes in your own development work that I have not noticed.

I would love to hear your perspective.

What do you think?

If you have a different way of looking at AI benchmarks, model evaluations, or where software development is heading, share it in the comments. We all come from different experiences, and I think that is where these conversations become much more interesting.

References

These are the main references behind the benchmark, evaluation, and capability points discussed in the article.

🤝 Let's Stay Connected

Place	Find me here
GitHub	building things → hemapriya-kanagala
LinkedIn	resources & updates → hemapriya-kanagala
X	random dev thoughts → @KanagalaHema

Dev Opportunity Radar #16: $15K AI Agent Hackathon, AI Education Fellowship, and AWS Student Rewards

Hemapriya Kanagala — Fri, 11 Sep 2026 13:59:00 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the AI for Education Fellowship, AWS Student Rewards, and the Multi-App AI Agent Hackathon, along with the Tech Interview Handbook as a resource worth checking out.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 AI for Education Fellowship
- 📌 AWS Student Rewards
- 📌 Multi-App AI Agent Hackathon
📚 Resources Worth Checking Out
- Tech Interview Handbook
🌟 Community Finds
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
AI for Education Fellowship	Worldwide Studios	Fellowship	Virtual (Global)	Applications reviewed on a rolling basis
AWS Student Rewards	AWS Builder Center	Student Rewards Program	Online (Global)	Open
Multi-App AI Agent Hackathon	Lemma × Comma Capital	AI Hackathon	Virtual (Global)	September 13

Resource Highlight

Resource	Why Check It Out
Tech Interview Handbook	A free collection of software engineering interview preparation material covering resumes, coding and algorithms, system design, behavioral interviews, interview best practices, and offer negotiation.

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

Opportunity	Organization	Type	Format	Deadline	Featured In
Agents for Humans Hackathon	AWS	Hackathon	Online (Mostly Global*)	September 15	Edition #14
MLH Global Hack Week: Data	Major League Hacking (MLH)	Hackathon	Online (Global)	September 17	Edition #14
HackerRank Orchestrate	HackerRank	AI Hackathon	Online (Global)	September 12	Edition #15
Claude Campus Ambassador Program	Anthropic	Campus Ambassador Program	Campus-based (Global)	September 12	Edition #15
Bharat Builds Tour	WeMakeDevs × AWS Builder Center	Student Hackathon Series	Online (India) / In Person	September 17–20	Edition #15

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 AI for Education Fellowship

Who it's for: Builders, engineers, educators, researchers, and anyone interested in building something at the intersection of AI and education. No coding experience or other prerequisites are required.

What stands out: The AI for Education Fellowship from Worldwide Studios is a 100-day virtual fellowship where you pick projects to work on, build and test them, and share what you're working on through demos every two weeks.

I wanted to include this because I like that it is focused on actually building something, rather than just learning about AI or education. You can explore an idea, build a first version, test it, and keep improving it throughout the fellowship.

The fellowship is fully virtual and you can join from anywhere in the world. Fellows can also work in teams if they want, and the community includes people from different backgrounds, including engineers, researchers, educators, and school owners.

Another thing worth knowing is the time commitment. The fellowship expects around 8–10 hours a week, with weekly squad syncs and bi-weekly demos. So this is something to think about if you're considering applying alongside work or school.

There are also examples of past fellows building and testing real projects during the fellowship, including AI tutoring and education-focused products.

Format: Virtual | Duration: 100 days

Dates: September 15 – December 15, 2026

Time Commitment: 8–10 hours per week

Cost: Free for accepted fellows

Eligibility: Anyone interested in building solutions in AI and education | No prerequisites

Application: Applications are reviewed as they come in

🔗 Learn More & Apply

📌 AWS Student Rewards

Who it's for: University students who want to learn cloud and AI, build projects, and earn AWS learning and certification benefits along the way.

What stands out: AWS Student Rewards on AWS Builder Center gives verified higher education students 12 months of AWS Skill Builder Premium, along with opportunities to earn AWS Credits and an AWS Foundational Certification exam voucher.

I wanted to include this because 12 months of premium learning access is a pretty significant amount of time. The Skill Builder access includes 900+ courses, hands-on labs, certification preparation, and other learning resources, so students can actually use it throughout the year rather than having access for only a short period.

There is also a reward system built around participating in the AWS Builder Center community. After verifying your student status and completing your profile, you get the Skill Builder reward. You can then earn badges through things like publishing articles, commenting, and maintaining engagement.

At 7 badges, you unlock $10 in AWS Credits, at 14 badges you get another $20, and at 21 badges you can unlock an AWS Foundational Certification exam voucher worth $100.

Another thing I liked is that AWS also has Student Builder Groups at universities in 60+ countries, so there is a community side to this as well if there is a group at your university.

One thing to check before signing up is the eligibility. You need to be 18 or older and enrolled at an accredited higher education institution, and your student status needs to be verified.

Benefits: 12 months of Skill Builder Premium + up to $30 in AWS Credits + AWS Foundational Certification exam voucher

Eligibility: Students 18+ enrolled at accredited higher education institutions worldwide

Cost: Free

🔗 Learn More & Get Student Rewards

📌 Multi-App AI Agent Hackathon

Who it's for: Developers, AI builders, and anyone who wants to build an AI agent that can actually take actions across different applications. You can participate solo or in a team of up to four.

What it's about: The Multi-App AI Agent Hackathon, hosted by Lemma and Comma Capital, is a one-day virtual hackathon where you build a multi-step AI agent connected to at least three external apps.

I wanted to include this one because the challenge is pretty clear: you're not just building a chatbot or showing what an AI agent could do. The idea is to build something that can actually take actions across multiple apps, and then show that it works.

The hackathon takes place on September 13, 2026, from 9 AM to 5 PM Pacific Time. You'll have most of the day to build, with judging and awards happening toward the end.

For the submission, you'll need a working project or repository, a two-minute demo, and a short system and reliability brief. The project is judged on technical execution, reliability and evaluation, usefulness, originality, and how clearly you can explain what you built.

There is also $15,000 in cash prizes, with $10,000 for first place, $4,000 for second, and $1,000 for third.

And probably the part many developers will care about: all three winning teams get a guaranteed interview with either Arga Labs or Lemma AI. Of course, that is an interview opportunity and not a job offer, but I think it makes the hackathon a little more interesting if you're also looking for opportunities to get your work in front of companies.

One thing to keep in mind is that this is a single-day hackathon, so you'll need to be comfortable building and making decisions fairly quickly. The challenge also has a specific requirement of connecting your agent to at least three external apps, so it's worth thinking about what you could build before the event starts.

Prize Pool: $15,000

Format: Virtual | Duration: 1 day

Date: September 13, 2026 | Time: 9:00 AM – 5:00 PM Pacific Time

Team Size: Solo or 2–4 people

What you need to build: A useful, multi-step AI agent connected to at least 3 external apps

Submission: Working project/repository + 2-minute demo + system and reliability brief

Judging: Technical execution, reliability & evaluation, usefulness, originality, and demo clarity

Hiring: Top 3 teams receive guaranteed interviews with Arga Labs or Lemma AI

🔗 Learn More & Register

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Tech Interview Handbook

If you're preparing for a software engineering interview and not sure what you should actually focus on, the Tech Interview Handbook is a free resource that puts a lot of the preparation material in one place.

It covers the whole interview process, starting from resume preparation and applying for jobs, then moving into coding and algorithms, system design, behavioral interviews, and finally offer negotiation.

I think the useful part here is that you don't have to use it only for coding practice. There are sections on things like how to prepare for coding interviews, common algorithm patterns, system design, behavioral questions, interview best practices, and questions you can ask the interviewer.

It also includes curated coding questions through the Blind 75 and Grind 75, along with explanations and tips for different algorithm topics. So if you don't want to randomly pick hundreds of problems to practice, you can use those lists as a starting point.

There are also sections on company interview formats, engineering levels, compensation, and choosing between offers, which can be useful once you get further into the process.

I wouldn't say you need to go through every section. Depending on where you are in your interview preparation, you can just pick the parts that are relevant to you. For example, if your coding preparation is already good, you could spend more time on system design or behavioral interviews.

So if you're preparing for interviews and looking for one place to get an idea of what to prepare and in what order, this is a resource I think is worth going through.

Format: Free online resource

Covers: Resume preparation, coding & algorithms, system design, behavioral interviews, interview best practices, offer negotiation, and more.

Also includes: Blind 75, Grind 75, curated practice questions, interview cheatsheets, and company interview information.

🔗 Explore the Tech Interview Handbook

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

We've had some really wonderful community finds over the last few weeks, and I honestly can't thank everyone enough for taking the time to find these opportunities, look through them, and send them my way. I know that takes time, and I appreciate every person who thinks of the Radar when they come across something they feel others might find useful 💙

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

So if you come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, please do send it my way. I love seeing what you find, and it helps make this series better for everyone.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

Dev Opportunity Radar #15: HackerRank Orchestrate, Claude Campus Ambassadors & the Bharat Builds Tour

Hemapriya Kanagala — Fri, 04 Sep 2026 17:37:52 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features HackerRank Orchestrate, the Claude Campus Ambassador Program, and the Bharat Builds Tour, along with the System Design Primer as a resource worth checking out.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features opportunities shared by L. Cordero (@earlgreyhot1701d) and Konark Sharma (@konark_13). Thank you both for sharing the NASA Space Apps Challenge and the Build with React Image Editor Challenge. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the Radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 HackerRank Orchestrate
- 📌 Claude Campus Ambassador Program
- 📌 Bharat Builds Tour
📚 Resources Worth Checking Out
- The System Design Primer
🌟 Community Finds
- The NASA Space Apps Challenge
- The Build with React Image Editor Challenge
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
HackerRank Orchestrate	HackerRank	AI Hackathon	Online (Global)	September 12
Claude Campus Ambassador Program	Anthropic	Campus Ambassador Program	Campus-based (Global)	September 12
Bharat Builds Tour	WeMakeDevs × AWS Builder Center	Student Hackathon Series	Online (India) / In Person	September 17–20

Resource Highlight

Resource	Why Check It Out
The System Design Primer	A free, open-source collection of system design topics, interview questions, real-world architectures, and resources for learning system design and preparing for system design interviews.

Community Finds

Shared By	Find	Type	Location / Format
L. Cordero (@earlgreyhot1701d)	NASA Space Apps Challenge	Global Hackathon	Global / 450+ Local Events
Konark Sharma (@konark_13)	Build with React Image Editor Challenge	Hackathon	Online

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the two most recent editions that are still accepting applications.

Opportunity	Organization	Type	Format	Deadline	Featured In
a16z Alpha Fellowship	a16z speedrun × EO Ventures	Fellowship	In Person	September 2026	Edition #13
RevenueCat Shipaton 2026	RevenueCat	Hackathon	Online	October 1	Edition #13
Kaggriculture AI Agent Competition	Kaggle	AI Competition	Online	September 30	Edition #13
Razorpay AI Buildathon	Razorpay	Buildathon	In Person (Bangalore, India)	September 5	Edition #14
Agents for Humans Hackathon	AWS	Hackathon	Online (Mostly Global*)	September 15	Edition #14
MLH Global Hack Week: Data	Major League Hacking (MLH)	Hackathon	Online (Global)	September 17	Edition #14

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 HackerRank Orchestrate

Who it's for: Developers, students, and builders who want to try building an AI agent in a short, focused challenge. It's open to anyone, and the challenge is solo.

What stands out: HackerRank Orchestrate is a 24-hour hackathon where you'll receive a real-world problem and build an agent-based solution for it. The challenge starts on September 12 at 6:00 PM IST, and you'll have 24 hours to build and submit your solution.

I wanted to include this one because it's a little different from the usual hackathons where you know the problem before you start. Here, the challenge is sent to you when the hackathon begins, so you'll need to figure out the architecture, retrieval strategy, model choice, prompts, and the rest of the approach yourself.

You can also use the tools you already work with, including Claude Code, Cursor, Codex, or Copilot. The submission includes your code, the agent's output, and your AI chat transcript, so they're looking at more than just the final result.

There's also an interesting part after submission: a 30-minute voice interview with an AI judge, where you'll walk through your architecture, decisions, tradeoffs, safety choices, and code. So if you're curious about what it feels like to have your AI project evaluated beyond a demo, this could be an interesting challenge to try.

There are cash prizes too, with $1,200 for first place and $300 for places 2–5, along with other rewards. The top three winners who build with Codex can also receive additional Codex credits.

One thing worth knowing before signing up is that this is a 24-hour challenge, and the problem itself isn't revealed until the start. So it may be more suited to people who enjoy working under a tight deadline rather than those looking for a longer, more relaxed hackathon.

Prize: $1,200 for 1st place | $300 for 2nd–5th place

Format: Online | Duration: 24 hours | Team: Solo

Challenge Starts: September 12, 2026 at 6:00 PM IST

What you need to build: An AI agent that solves a real-world problem provided when the challenge begins.

Evaluation: Code, agent output, AI chat transcript, and a 30-minute AI judge interview.

Eligibility: Anyone can participate | No prior AI experience required

🔗 Learn More & Register

📌 Claude Campus Ambassador Program

Who it's for: Students who want to help other students learn about and use AI on their campus. There are separate tracks for undergraduates, master's students, and PhD students/postdocs, with different responsibilities for each.

What stands out: The Claude Campus Ambassador program from Anthropic is back for the 2026–27 school year, and applications are currently open.

I wanted to include this because it's a little different from the usual student programs I come across. Instead of being focused mainly on learning or completing a project, you're helping build an AI community on your campus and organizing activities for other students.

For undergraduate students, the program involves leading a Claude Builder Club, organizing technical and non-technical workshops, running hackathons or demo nights, and sharing resources with members. Selected ambassadors receive a $3,600 cash stipend.

For master's students, there's a Claude Campus Conversations track focused more on discussions around AI and its impact on education and society. You'd organize discussion groups and events, while also helping students share and discuss the AI projects they're building.

For PhD students and postdocs, the Claude Science track is focused on researchers using AI in their work. This includes leading workshops for other researchers, organizing Claude for Science sessions, and engaging with Anthropic's science teams.

All three tracks also give ambassadors opportunities to participate in workshops led by Anthropic, connect with other ambassadors, and share feedback with Anthropic's teams.

A few things are worth knowing before applying. The program runs for the full school year, from September 2026 through June 2027, so this is more of an ongoing commitment than a short-term program. You also need to be 18 or older and have work authorization in your country of study.

The program is open to students anywhere in the world, although the initial pilot for the graduate and PhD tracks may select from a smaller group of schools. You don't need to be an AI expert, but Anthropic does expect some level of AI fluency and familiarity with Claude.

Stipend: $3,600

Format: Campus-based | Program Duration: September 2026 – June 2027

Tracks: Claude Builder Club (Undergraduate), Claude Campus Conversations (Master's), and Claude Science (PhD/Postdocs)

Eligibility: Students worldwide | Age: 18+

Application Deadline: September 12, 2026

Cost: Free to apply and participate

🔗 Learn More & Apply

📌 Bharat Builds Tour

Who it's for: University students in India who want to build projects through a series of online and in-person hackathons. You must be 18 or older to register on your own.

What stands out: The Bharat Builds Tour, organized by WeMakeDevs in collaboration with AWS Builder Center, is a six-stop hackathon series where students can participate online from anywhere in India. Each stop has its own hackathon, and you can take part in as many of the six as you'd like.

I wanted to include this one even though it's geographically limited because there is a hiring component that I don't come across very often in student hackathons.

For students graduating in 2027 or 2028, some of the top projects can be eligible for a fast-track interview opportunity with Amazon's University Talent Acquisition Team for available internship and full-time roles. The fast-track means the usual screening round is skipped and the project is used in its place.

That doesn't mean winning a hackathon automatically gets you an Amazon interview, though. The rules say that being shortlisted for a prize and being put forward for an interview are separate decisions, and the interview opportunity is also subject to Amazon's eligibility requirements, available roles, and business needs. So I'd definitely read the terms before counting on the hiring part.

The tour itself includes six hackathons, with the first one, First Commit, taking place September 17–20. Every stop has an online option for students across India, while the city hosting that particular stop also has limited in-person seats.

You can participate solo or in teams of up to four, and you don't need previous cloud experience. Each project does need to use AWS, and submissions include a public repository, a short demo video, and a write-up explaining the project and where AWS is used.

There are also prizes across the tour, with a total prize value of ₹1 crore in gadgets and cloud credits. Participants can also get access to AWS learning resources, credits, a certification voucher, mentors, and a WeMakeDevs certificate.

One thing worth knowing is that one registration covers all six stops, but you still need to check in to each hackathon as it opens. You'll also need an AWS Builder Center profile and verified student status before your entry can be scored.

Prize Pool: Prizes worth ₹1 crore across six hackathons

Format: Online across India + limited in-person seats

First Hackathon: September 17–20, 2026

Team Size: Solo or up to 4 people

Eligibility: University students in India, 18+

Hiring: Top projects from eligible 2027 and 2028 students may receive a fast-track interview opportunity with Amazon

Cost: Free

🔗 Learn More & Register

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

The System Design Primer

If system design is something you're trying to learn, this is one resource I came across that is worth keeping around.

The System Design Primer is a free, open-source GitHub repository with a collection of system design topics, interview questions, real-world architectures, and other resources. It also has a study guide that gives you an idea of what to focus on depending on how much time you have.

It covers things like scalability, latency, availability, consistency, DNS, CDNs, load balancers, databases, caching, microservices, and distributed systems.

What I liked here is that it also has actual system design questions to work through, such as Pastebin, Twitter, a web crawler, a social network, and scaling a system to millions of users on AWS.

There is also a section on how to approach a system design interview, starting with requirements and constraints, then moving into the high-level design, core components, scaling, and trade-offs.

So if you're just getting started with system design, or preparing for interviews and don't know what topics to go through first, this could be a good place to start.

Format: Free, open-source GitHub repository

Covers: System design, scalability, databases, caching, distributed systems, networking, microservices, and more.

Also includes: System design interview questions, real-world architectures, and Anki flashcards.

🔗 Explore the System Design Primer

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

The NASA Space Apps Challenge

Shared by L. Cordero (@earlgreyhot1701d).

L. Cordero shared the NASA Space Apps Challenge, a global hackathon happening on November 14–15, 2026, with 450+ local events to choose from when you register.

The challenge brings people together to work on real-world problems using NASA data and Space Agency Partner data. The challenges can cover things like software development, storytelling, astrophysics, space exploration, agriculture, and more.

I wanted to include this because I like the community side of this one. You can choose a local event and meet other people interested in technology, science, and space, and it is open to people from different backgrounds and skill levels.

Another thing L. Cordero pointed out is that this is far enough out to plan around. There is still some time to find teammates, look at the local events, and think about what you might want to build.

L. Cordero also mentioned the idea of putting together a DEV team for this. I won't be able to join this one myself with everything going on right now, but maybe some of you would be interested in putting a team together.

Format: Global hackathon with local events

Dates: November 14–15, 2026

Local Events: 450+

Who can participate: All ages, skill levels, and professional backgrounds

🔗 Find a Local Event & Learn More

💡 Community Note: If you're interested in joining a DEV team for this, feel free to leave a comment below and connect with others who are interested. Maybe we can get a few people together for this one!

The Build with React Image Editor Challenge

Shared by Konark Sharma (@konark_13).

Konark shared the Build with React Image Editor Challenge from Unlayer, where you build a GTA VI-inspired experience using React Image Editor.

The idea is to build something where users can edit or customize at least one visual, use React Image Editor as a core part of the project, and then publish the source code in a public GitHub repository and deploy the project.

I wanted to include this because it's a little different from the usual hackathons I come across. The GTA VI theme gives you a starting point, but there is still quite a bit of room for what you can actually build around it.

There is $1,000 in prizes, with $500 for first place, $300 for second, and $200 for third. Projects are judged on creativity, visual execution, use of React Image Editor, and the overall experience.

Format: Online

Deadline: September 24, 2026 at 11:59 PM UTC

Prize Pool: $1,000

What you need to build: A GTA VI-inspired experience using React Image Editor, with at least one visual that users can edit or customize.

Submission: Public GitHub repository + deployed project

🔗 Learn More & Submit

Thank you to L. Cordero (@earlgreyhot1701d) and Konark Sharma (@konark_13) for thinking of the Radar and sharing these with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

I Built My First AWS Agent Workflow, and the Hardest Part Was Getting It to Stop Assuming Things

Hemapriya Kanagala — Thu, 03 Sep 2026 15:24:18 +0000

TL;DR

I recently finished a project from Udacity's Future AWS Agent Engineer Nanodegree Program, which I was able to take through the AWS AI & ML Scholarship.

I built a customer support agent using Amazon Bedrock AgentCore, AgentCore Gateway, AWS Lambda, DynamoDB, and an FAQ. The agent had to understand whether a customer was reporting a bug, asking a question that could be answered from the FAQ, or asking for something that needed human support.

I finished it in about two days and passed on my first attempt with a correctness score of 0.83. I did not have enough time to go back and refine it or run another evaluation round because there was a lot happening in my life at the time.

The part I learned the most from was actually one of my evaluation failures. For a bug report, I had told the agent that it needed 3 things before creating a ticket: a description of the problem, steps to reproduce it, and the environment where it happened. But in 2 test cases, the model understood what the customer was talking about and created the ticket even though one of those required pieces of information was missing.

That made me realize something I had not thought about enough before: understanding a request is not the same as having enough information to safely take an action.

That is the main lesson I took from this project. Building an agent is not only about making the model capable of answering. It is also about being clear about when it should act, when it should ask for more information, and when it should stop and hand the request to a person.

I know there are already many great articles about AWS, AgentCore, and AI agents, and I have learned from many of them myself. I am still learning too, so this is simply my experience of building this project, what confused me, what the evaluation showed me, and what I would do differently next time.

I Want to Start With Something Honest
What I Built and How the Pieces Fit Together
- So What Does Each AWS Service Actually Do Here?
The Prompt Was Doing More Than I Expected
- The Prompt Was Setting the Rules
- I Also Had to Think About What the Customer Might Try
The Evaluation Showed Me Where I Was Wrong
- The Two Failures
What I Would Change and What I Learned
- Seeing the Backend Made the Agent Make More Sense
If You Are Also Starting Out
I Would Love to Hear From You
A Small Thank You
My Final Takeaway
🤝 Let's Stay Connected

I Want to Start With Something Honest

I finished another project recently, and this one was a little different for me. It was part of the Future AWS Agent Engineer Nanodegree Program from Udacity, which I was able to take through the AWS AI & ML Scholarship.

I finished the project in roughly 2 days. I was working around a lot of other things happening in my life, so I was trying to make the most of the time I had and keep moving.

I passed the project on my first attempt, which I was really happy about, but I also knew I had not spent the time I normally would on refining it.

If I had more time, I would have gone back, looked at the evaluation results more carefully, improved the prompt, added more test cases, and run another evaluation. I did not do that this time. The final correctness score was 0.83.

At first, I felt like I should probably wait until I had a higher score before writing about the project. Then I thought about what I had actually learned from it. The 2 things that did not work perfectly were some of the most useful parts of the project for me. That is what I wanted to share.

There are already a lot of really good articles about building agents, AWS services, and AgentCore. I have learned from many of them myself. This is not meant to be another article that tries to explain everything about agent development. Instead, I wanted to share what the project looked like from my side as someone who is still learning.

Because when you are a beginner, sometimes the hardest part is not the code. It is understanding what all the pieces are actually doing and why they need to be there.

What I Built and How the Pieces Fit Together

The project was a fictional customer support system for an online shop. A customer could send a message, and the agent had to understand what the customer needed and decide what to do next.

There were three possible behaviors. The first was BUG_REPORT. If a customer said something like, “The checkout page crashes every time I click Pay,” the agent needed to recognize that as a technical problem. But recognizing it as a bug was not enough to create a ticket. Before doing that, the agent needed to collect three specific pieces of information: description, steps to reproduce, and environment.

The second was PLATFORM_QUESTION. These were questions about things such as orders, shipping, returns, refunds, payments, products, accounts, and privacy. For these questions, the agent had an FAQ that it was expected to use instead of making up an answer from what it generally knew.

The third was OTHER_REQUEST. If the request was not a technical bug and could not be answered using the FAQ, the agent needed to direct the customer to human support.

Once I stopped looking at each AWS service separately and looked at how they worked together, the architecture became much easier for me to understand.

The flow was essentially:

This is the architecture I built for the project. The routing was handled through the system prompt rather than through a separate classifier or condition node.

So what does each AWS service actually do here?

If you are new to AWS, the service names can make a project like this look much more complicated than it really is. What helped me was thinking about what each piece was responsible for instead of trying to understand every AWS service at once.

Amazon Bedrock AgentCore Managed Harness was the environment around the agent. In simple terms, it gave me the structure needed to run the agent with its model, instructions, and tools without having to build all of that setup myself.

The model was responsible for understanding the customer's message and deciding what to do next based on the instructions it was given. It was the part doing the actual language understanding and decision-making.

The system prompt defined how the agent was supposed to behave. This was especially important in my project because the routing logic was written into the prompt. The prompt told the model how to recognize the 3 types of requests and what rules to follow for each one.

AgentCore Gateway connected the agent to the backend tool. The model was not directly changing the database. Instead, once the required information had been collected, the agent could call the tool through the Gateway.

AWS Lambda handled the actual backend operation. When the agent called the tool with the required information, Lambda processed the request and created the bug report.

Amazon DynamoDB stored that bug report. This was also a useful thing for me to understand because it showed that the agent's response was not just text. If the agent returned a ticket ID, there was an actual record stored in DynamoDB behind that response.

And the FAQ was the controlled source of information for platform questions. The agent was expected to use the information provided there instead of relying on general knowledge when answering those questions.

Once I looked at it this way, the architecture stopped feeling like a collection of unfamiliar AWS names. The model understood the customer's message, the prompt told it what rules to follow, the Gateway connected it to an action, Lambda performed that action, and DynamoDB stored the result. For platform questions, the FAQ provided the information the agent was allowed to use, while requests outside those boundaries went to human support.

That simple way of looking at the system helped me more than memorizing what each AWS service does on its own.

One thing that confused me at first was the word “classifier.” When I first saw that requirement, I imagined there would be a separate component whose only job was to classify the customer's message and then send it to the right part of the system.

That was not what I built.

In my implementation, the routing logic was part of the system prompt. The model was instructed to choose one of the 3 routes and then follow the rules for that route.

Understanding this changed how I looked at the prompt. I initially thought of the prompt mostly as instructions for how the agent should respond. In this project, it was doing much more than that. It was also defining what information the agent needed, when it was allowed to take an action, what information it could use to answer a question, and when it needed to stop and involve human support.

The Prompt Was Doing More Than I Expected

Before this project, I mostly thought about a system prompt as a way of telling the model how to behave. Something like:

“You are a helpful customer support assistant.”

That is useful, but it is not enough when the model is also making decisions and using tools. In this project, the prompt was not just telling the agent how to respond. It was also telling it what it was allowed to do and what information it needed before doing it.

The Prompt Was Setting the Rules

For a BUG_REPORT, the prompt explained what counted as a bug and what information was required before a ticket could be created. The 3 required fields were description, steps to reproduce, and environment. If something was missing, the agent had to ask for that information one field at a time. It was not supposed to call the tool until all 3 were available.

The prompt also defined the other 2 routes. A PLATFORM_QUESTION had to be answered using the provided FAQ, while an OTHER_REQUEST had to be directed to human support.

The FAQ part was interesting to me because it showed why an agent sometimes needs to be told what information it can and cannot use.

Suppose a customer asks:

“How long do I have to return an item?”

The FAQ says that most items can be returned within 30 days of delivery, as long as they are unused and in their original packaging, unless the item arrived defective. So the agent can answer that using the information it was given.

But suppose someone asks:

“Do you offer a student discount?”

If the FAQ does not mention student discounts, the agent should not simply make up an answer because the model might have seen student discounts mentioned somewhere else. It should send the customer to human support.

That made something very clear to me: a model can know a lot, but that does not mean it knows your company's policy.

The same applies to things like refunds, shipping, payment policies, account rules, and other information where giving a confident but incorrect answer could cause a real problem. For this project, the FAQ was a controlled source of information, and the prompt made that boundary clear.

I Also Had to Think About What the Customer Might Try

The prompt had another job that I had not thought much about before this project: dealing with prompt injection.

One of the test cases essentially told the agent to ignore its instructions, reveal the system prompt, and create a bug ticket without collecting the required information. The agent did not follow those instructions. It continued following the rules defined in the system prompt.

I found this useful because it made me think beyond the simple test cases I naturally wrote while building the project. When we test an agent ourselves, we usually give it the kind of message we expect a customer to send. Real users can leave information out, ask something completely unrelated, or deliberately try to change how the agent behaves.

So the prompt cannot only describe the happy path. It also needs to make the boundaries clear.

At the same time, passing this one prompt-injection test does not mean the agent is protected against every possible injection. It only showed me that this particular test case was handled according to the rules I had defined.

That was another lesson from the project for me. Giving an agent more capabilities is only part of the work. You also have to be clear about when those capabilities should and should not be used.

The Evaluation Showed Me Where I Was Wrong

This was probably the most useful part of the project for me.

Getting the agent to respond correctly in a few manual tests is one thing. Evaluating it with deliberately incomplete and tricky inputs is another. For the project, I created a test dataset covering the different routes, including complete bug reports, bug reports with missing information, FAQ questions, questions that were not covered by the FAQ, requests to speak to a real person, and a prompt-injection case.

The final correctness score was 0.83. Most of the tests behaved the way I wanted. Complete bug reports worked, the missing-environment case worked, the FAQ questions worked, unsupported questions were handed off, requests for human support were handled correctly, and the prompt-injection case also passed.

But two tests failed, and both failures taught me basically the same thing.

The Two Failures

In the first case, the customer said that product images were not loading on category pages and mentioned Safari on an iPhone. The problem description was there, and the environment was there, but the customer had not explicitly provided reproduction steps. The agent still created a ticket.

In the second case, the customer gave reproduction steps and a Chrome/Windows environment, but did not provide a clear description of the actual problem. Again, the agent created a ticket.

At first, I thought the model had simply misunderstood my prompt. But when I looked at the two cases more carefully, I realized that the model probably understood what the customers were talking about. The problem was that it treated that understanding as enough information to take the next action.

From a normal human conversation perspective, that is not necessarily a bad thing. If someone says:

“The product images are not loading on Safari on my iPhone.”

a person can probably understand what is going wrong and know what the customer is referring to.

But that was not the rule I had given the agent. The agent was not supposed to decide, “I understand the problem, so I can create the ticket.” It was supposed to check whether all 3 required fields were explicitly available before creating the ticket.

That is a much stricter requirement.

This was probably the biggest lesson I got from the project: understanding something is not the same as having enough information to safely perform an action.

That matters even more when an agent starts calling tools. If the same kind of assumption happened in a system that creates a refund, changes an account, submits an application, or updates a database, the model might make a very reasonable guess, but that guess could still lead to the wrong action.

So for me, the important question changed from:

“Does the model understand what the customer means?”

to:

“Does the model have everything it is required to have before it takes this action?”

That is something I would pay much more attention to if I built another agent.

Looking back, I actually found the individual evaluation results more useful than the overall score. The 0.83 tells me how the agent performed overall, but looking at the individual cases showed me exactly where the behavior broke down.

For me, those two failed cases were more useful than simply knowing that the score was not 1.0. They showed me a specific problem in the way I had defined the agent's behavior, and more importantly, they gave me something concrete that I could improve.

What I Would Change and What I Learned

If I had another day to work on the project, I would start with those two failed cases.

I would make the prompt more explicit about what counts as a field being present, especially the difference between a description and reproduction steps. For example, “The images are not loading” is a description of the problem. “Open the shop, go to a category, and tap any product” is a sequence of steps someone could follow to reproduce it.

I would also make it clearer that the agent should not fill in a missing field just because the rest of the message makes the situation easy to understand. If the customer has not actually provided the reproduction steps, the agent should ask for them instead of deciding that it already knows what they probably are.

Then I would add more tests around the same problem. What happens when the customer provides two fields in one sentence? What happens when the third field comes in a later message? What happens when the customer uses vague language? What happens when the information could reasonably fit into more than one field?

Those are the cases I would want to explore next because they are the kinds of situations where an agent can appear to be doing well while still making assumptions.

Seeing the Backend Made the Agent Make More Sense

I also learned something from the backend side of the project. Before building it, I mostly thought of an agent as something that takes a message and gives you an answer. Working with the tool and the backend made me see what happens when the agent is expected to actually do something.

The flow in my project was:

Customer information
        ↓
Agent checks whether enough information exists
        ↓
AgentCore Gateway
        ↓
Lambda
        ↓
DynamoDB
        ↓
Ticket ID returned
        ↓
Customer receives the ticket ID

The important part for me was that the model's response was not the action itself. When the agent said that a ticket had been created, there needed to be an actual operation behind that response. The tool call went through AgentCore Gateway, Lambda handled the backend operation, and DynamoDB stored the ticket. The ticket ID returned from that process was then given back to the customer.

That made the difference between a regular chatbot and an agent workflow much clearer to me. A chatbot can mainly respond to what you say. An agent can also be given tools that allow it to take actions outside the conversation.

But once you give a model that ability, you also have to be careful about what happens when things do not go as expected.

The model should not say that a ticket was created if the tool failed. It should not invent a ticket ID if one was not returned. And if the FAQ does not contain an answer, it should not create one just because it can produce a convincing response.

These might sound like obvious rules, but actually building the workflow made me understand why they matter. The model is only one part of the application. The prompt defines what it should do, the tools perform actions, the backend systems store or change information, and the application needs to make sure those pieces agree with each other.

That changed how I think about agents. I no longer see the model as the whole application. I see it as one part of a larger system where the instructions, tools, data, and backend all have to work together.

If You Are Also Starting Out

I am writing this part especially for beginners because I am one too.

I did not start this project knowing exactly how all of these services worked together. A lot of the environments I had worked with before already had things set up for me. You work inside the provided environment, use the tools that are already there, and gradually learn how everything fits together.

This project made me look underneath that layer. I had to understand what the AgentCore harness was doing, why the Gateway was needed, how Lambda connected the tool to the backend, why DynamoDB was there, and how the evaluation could show me that something I thought was working was actually not behaving according to the requirements.

That was probably the biggest difference for me.

When you build a normal application, it can be easy to think about it as:

Input
   ↓
Code
   ↓
Output

With an agent, the model is also making decisions about what should happen next. So I had to start thinking about things I had not really considered before. What is the model allowed to decide? What information does it need before it can act? What should it do when information is missing? What should it do when it does not know the answer? What happens when someone tries to make it ignore its instructions? And what should happen if the tool itself fails?

Those questions are just as important as connecting the AWS services. That was something I did not fully appreciate before working on the project.

I also realized that I did not need to understand every AWS service deeply before I could start building. What helped me was asking a much simpler question whenever I came across something unfamiliar:

What problem is this service solving in my application?

For DynamoDB, I did not need to understand everything about the database service. I needed to understand that my application needed somewhere to store the bug report.

For Lambda, I did not need to know every feature it provides. I needed to understand that something had to actually execute the backend operation when the agent called the tool.

For AgentCore Gateway, I did not need to know every capability it has. I needed to understand that it provided the connection between the agent and the tool.

Thinking about the services this way made the architecture much less overwhelming for me. Instead of trying to learn a huge list of AWS services first and then figuring out where they fit, I could start with what my application needed and learn the service that solved that particular problem.

I am also glad I started this project even though I did not finish it in the ideal way I normally would have. I completed it in about 2 days, passed on my first attempt, and then had to move on instead of spending another round going back through everything and trying to improve the 0.83 score.

Of course, part of me would have liked to see what I could have changed and whether another evaluation would have given me a better result. But I am also glad I did not wait until I felt completely ready before starting.

There were moments when I was learning something and immediately had to figure out how to use it in the project. That was uncomfortable sometimes, especially when I was still trying to understand what a service actually did, but it also made the concepts much easier for me to remember.

I saw where AgentCore Gateway fit into an actual workflow instead of only reading about it. I saw Lambda handle the backend operation and DynamoDB store the result. And I saw the evaluation catch something I had missed myself.

That is probably what I will remember from this project more than the final score. I started without knowing everything, built something with what I understood, found places where it could be better, and learned from those places.

I Would Love to Hear From You

I am still learning, so I would love to hear from people at different stages of this.

If you have tips, advice, lessons from your own projects, mistakes that taught you something, or anything you think might help someone who is just starting out, please share it in the comments.

One of the things I love about DEV is that people come from such different experiences. So if you have something to add, even if it seems small, please share it. Maybe something you learned along the way can help someone else who is just starting.

A Small Thank You

I also want to thank Udacity and the people involved in the Future AWS Agent Engineer Nanodegree Program. Having the project structure, learning material, and live sessions made it much easier for me to keep going when I was confused about something.

The live sessions were especially useful because sometimes you can read something several times and still not understand how it fits into the actual project. Hearing someone walk through it and then trying it yourself makes a difference.

I am still learning, but having a project where I could actually put these concepts together helped me understand them in a way that reading about them alone probably would not have.

My Final Takeaway

If I had to take one thing from this project, it would be this:

A good agent is not only one that knows what to do. It also needs to know when it has enough information to do it.

That sounds simple now, but I understood it differently after seeing the evaluation catch my own assumption.

The model can understand the customer. The prompt can give it instructions. The tools can let it take actions. The backend can store the result. But all of those pieces still need clear rules about when the next step is actually allowed to happen.

That is what I found most interesting about building this project.

I started by thinking I was mainly learning how to put AWS services together. I ended up learning much more about how important the boundaries around an agent are.

And honestly, that is probably a better lesson for me to take into the next project.

🤝 Let's Stay Connected

Place	Find me here
GitHub	building things → hemapriya-kanagala
LinkedIn	resources & updates → hemapriya-kanagala
X	random dev thoughts → @KanagalaHema

Dev Opportunity Radar #14: MLH Global Hack Week, $40K Agents for Humans Hackathon & Razorpay AI Buildathon

Hemapriya Kanagala — Fri, 28 Aug 2026 13:24:07 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the Razorpay AI Buildathon, Agents for Humans Hackathon, and MLH Global Hack Week: Data, along with Andrej Karpathy's Deep Dive into LLMs like ChatGPT as a resource worth checking out.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

If you've discovered something through the radar, I'd love to hear about it. Whether you applied to an opportunity, joined a community, completed a program, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, or event that deserves more attention, feel free to share it in the comments.

If I feature one of your Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Razorpay AI Buildathon
- 📌 Agents for Humans Hackathon
- 📌 MLH Global Hack Week: Data
📚 Resources Worth Checking Out
- Deep Dive into LLMs like ChatGPT
🌟 Community Finds
💙 Reader Updates
💙 A Little Personal Note
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
Razorpay AI Buildathon	Razorpay	Buildathon	In Person (Bangalore, India)	September 5
Agents for Humans Hackathon	AWS	Hackathon	Online (Mostly Global*)	September 15
MLH Global Hack Week: Data	Major League Hacking (MLH)	Hackathon	Online (Global)	September 17

Agents for Humans Hackathon: Open internationally, but some countries and territories are excluded. Check the full eligibility rules before participating.

Resource Highlight

Resource	Why Check It Out
Deep Dive into LLMs like ChatGPT	A free 3+ hour deep dive by Andrej Karpathy covering how LLMs work, from tokenization and neural network internals to training, inference, hallucinations, tool use, and reinforcement learning.

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from the previous two editions that are still accepting applications.

Since this is already Edition #14, the list of opportunities covered across the earlier editions is starting to get quite long. I used to keep all the still-open opportunities here, but I don't think repeating the full list every week makes the editions easier to read.

So from this edition onward, I'll keep this section to the two most recent editions. This way, you can still get a quick reminder about recent opportunities without making the weekly edition too long.

If you're looking for opportunities from earlier editions, they're still available on the Dev Opportunity Radar website, where you can browse the full list and use the filters to find what you're looking for.

I've already covered the opportunities below in detail, so I won't repeat everything here. If any of them catch your attention, you can find the full overview, eligibility details, and application links in the original edition.

Opportunity	Organization	Type	Format	Deadline	Featured In
Magnificent Grants	Magnificent Grants	Grant & Fellowship	Hybrid	Rolling	Edition #12
AI Societal Impact Lab Autumn 2026 Fellowship	AI Societal Impact Lab	Fellowship	Remote	September 4	Edition #12
a16z Alpha Fellowship	a16z speedrun × EO Ventures	Fellowship	In Person	September 2026	Edition #13
RevenueCat Shipaton 2026	RevenueCat	Hackathon	Online	October 1	Edition #13
Kaggriculture AI Agent Competition	Kaggle	AI Competition	Online	September 30	Edition #13

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Razorpay AI Buildathon

Who it's for: Students interested in AI and software engineering who are able to work in person in Bangalore from September 2026.

What stands out: The Razorpay AI Buildathon is a student-only program where you build an AI project and use that project to apply for a 6 or 12-month AI Builder Internship at Razorpay.

There are five tracks you can choose from: AI Growth & Agentic Commerce, AI Risk Manager, AI Revenue Recovery, AI Finance Controller, and an Open Track if you want to build something outside the listed ideas.

I wanted to include this because I don't come across many opportunities where the project you build is actually part of the hiring process. Here, Razorpay is looking at how you think, what you build, how well it works, and even how you deal with things when they break.

Another thing I liked is that they don't make the application all about your resume. You'll need to submit a public GitHub repository, a five-minute pitch video, explain what your project solves, and talk about what broke and how you fixed it. There is also no aptitude test or group discussion. Shortlisted builders go directly to a panel.

The internship itself is also quite substantial. Selected students can choose between 6 or 12 months and receive a ₹75,000 monthly stipend while working in person in Bangalore.

I know many people reading the radar are looking for opportunities that are accessible globally, so this one is definitely more location-specific. I'm still sharing it because I think the combination of building something real + getting evaluated on your work + potentially turning that into a paid AI engineering internship makes this worth knowing about, especially for students in India who can be in Bangalore.

Internship: 6 or 12 months | Stipend: ₹75,000/month

Format: In person, Bangalore

Application Deadline: September 5, 2026

Application Requirements: Resume, project details, public GitHub repository, five-minute pitch video, and application questions.

🔗 Learn More & Apply

📌 Agents for Humans Hackathon

Who it's for: Developers, students, and builders who want to build an AI agent that can handle repetitive tasks for people. You can participate solo or with a team, and the hackathon is open to participants from around the world, with some countries and territories excluded.

What stands out: The Agents for Humans Hackathon, organized by AWS, is a six-week online hackathon where you build a new AI agent using the Strands Agents SDK. The idea is pretty simple: build something that can take care of a real task that people normally have to do themselves.

There are three tracks to choose from: Everyday Agents, for things like home, money, errands, or family tasks; Professional Agents, for repetitive work tasks; and Good Neighbor Agents, for agents that help groups such as nonprofits, schools, libraries, or local organizations.

I wanted to include this because I think the theme leaves quite a lot of room to build something you would actually want to use. You're not given one specific problem to solve, so you can start with a repetitive task you've personally dealt with and think about whether an agent could handle it.

Another thing I like is that the submission isn't just about showing an idea. You need to build a working agent, make your code publicly available, include an architecture diagram, and record a short demo showing it working. So if you're looking for a project that can also become something useful for your portfolio, this could be an interesting one to try.

The hackathon also provides $50 in AWS credits to help with build costs, and there is a total of $40,000 in cash prizes across the three tracks.

Prize Pool: $40,000

Format: Online | Build Period: Six weeks

Submission Deadline: September 15, 2026

Team Size: Solo or teams

What you need to build: A new AI agent using the Strands Agents SDK that handles a real, repetitive task.

Tracks: Everyday Agents, Professional Agents, and Good Neighbor Agents.

Additional Support: $50 in AWS credits are available for the build.

🔗 Learn More & Join the Hackathon

📌 MLH Global Hack Week: Data

Who it's for: Students, developers, and anyone interested in learning more about data, whether you're just getting started or already have some experience.

What stands out: Global Hack Week: Data is a free, week-long online event from Major League Hacking (MLH) where you can learn, build projects, complete challenges, and meet other people from the MLH community.

The week includes live workshops and technical challenges covering topics like SQL, databases, and data visualization. There are also smaller community events on Discord where you can meet other participants, work together, and take a break between challenges.

I wanted to include this because sometimes learning data skills can feel like you need to take a full course before you can actually start building something. This is a more relaxed way to spend a week exploring the area, learning something new, and actually putting it into practice.

Another thing I like is that you don't have to complete every challenge or already know what you're doing. Some challenges are simple, like connecting with other members of the community, while others involve building projects and creating demos. You can choose what you want to take on depending on your interests and experience.

And this one is actually global. It's completely free and open to participants from anywhere in the world, which is always something I look for when going through opportunities for the radar.

Event: September 11–17, 2026

Format: Fully online | Cost: Free

Topics: SQL, databases, data visualization, and other data-related technical skills.

🔗 Learn More & Register

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

Deep Dive into LLMs like ChatGPT

Who it's for: Developers, students, AI enthusiasts, and anyone who wants to understand how large language models like ChatGPT actually work. You don't need to be an AI researcher to follow it, but some basic technical knowledge will make it easier to get through the deeper sections.

What stands out: This is a free 3+ hour deep dive into LLMs by Andrej Karpathy, covering much more than just how to use ChatGPT. He walks through the full training stack, starting with internet data and tokenization, then moving into neural network internals, inference, pretraining, post-training, hallucinations, tool use, reinforcement learning, and how models actually "think."

I wanted to include this because there are so many resources about using AI tools, but fewer that actually help you understand what's happening underneath. If you've ever wondered how a large language model goes from a huge amount of text to something that can answer questions, write code, or use tools, this is a really good place to start.

Another thing I liked is that Karpathy doesn't only explain the technical side. He also talks about some of the strange ways LLMs behave, including hallucinations, tokenization problems, "jagged intelligence," knowledge and working memory, and why models can be surprisingly good at some things while struggling with others. That makes the video useful even if you're more interested in building with AI than training models yourself.

Format: Free recorded video | Length: Approximately 3 hours 22 minutes

Level: Beginner-friendly, with deeper technical sections

Topics Covered: LLM pretraining, tokenization, neural networks, inference, post-training, hallucinations, tool use, reinforcement learning, RLHF, DeepSeek-R1, and practical use of LLMs.

🔗 Watch the Video

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

This week, we weren't able to include a community find, but that's completely okay. Hopefully we'll have more wonderful finds from the community as we keep going.

If I feature something you shared in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

💙 A Little Personal Note

I know I've been a little quieter around the community over the last couple of weeks. I haven't been as active as I normally am, and I know I haven't been able to support everyone's posts and work in the same way I usually would.

I've been going through a difficult phase personally, with some things happening in my family, and I've needed to take a little step back. I haven't really talked about it here because I didn't quite know how to put it into words. But I wanted to share a little bit of it because so many of you have continued to support me and the Radar.

I'm sharing this here mainly because I don't want my being quiet to be misunderstood. This community means a lot to me, and I know many of you regularly support what I share. I don't want someone to think that I saw their work and chose not to support it, or that I don't care, when that isn't the case.

Even though I've been quieter, I'm still trying to keep up with everyone. Whenever I come online, especially when I'm working on the Radar on Fridays or whenever I have the energy during the week, I try to go through the posts and see what everyone has been working on. I may not always have the energy to leave a comment or reaction on everything, but I am still looking, reading, and cheering for you from my side.

So please don't take it personally if I haven't reacted to or commented on something you've shared. If I missed your post, there's a very good chance I simply missed it, not that I don't care about what you're doing. I know I may have missed quite a few things recently, and I'm genuinely sorry for that.

I also know that many of you continue to give me so much support, and I wish I could always give the same energy back. Right now, I just can't always be as present as I'd like to be. I'm doing what I can with what I have, and I hope you'll understand that.

I don't want this difficult phase to stop me from doing the things I care about either. The Radar is one of those things. Even if I'm somewhat quiet around the community for a while, I'll keep coming back every Friday with a new edition and keep doing my best to find and share opportunities that might help someone.

Thank you to everyone who has continued to support me, encourage me, read the editions, share opportunities, leave comments, and simply be here. I notice it, even when I don't always have the energy to respond properly. It means a lot to me 💙

And if I've missed your post recently, please know that I haven't stopped caring. I might have just missed it. I hope you guys understand.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

Dev Opportunity Radar #13: a16z Alpha, a $740K Hackathon, and an AI Agent Competition

Hemapriya Kanagala — Fri, 21 Aug 2026 17:06:59 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the a16z Alpha Fellowship, RevenueCat Shipaton 2026, and the Kaggriculture AI Agent Competition, along with IBM SkillsBuild as a resource worth checking out.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features an opportunity shared by Konark Sharma (@konark_13). Thank you, Konark, for sharing the Agent Harness Hackathon. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 a16z Alpha Fellowship
- 📌 RevenueCat Shipaton 2026
- 📌 Kaggriculture AI Agent Competition
📚 Resources Worth Checking Out
- IBM SkillsBuild
🌟 Community Finds
- The Agent Harness Hackathon
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Location / Format	Deadline
a16z Alpha Fellowship	a16z speedrun × EO Ventures	Fellowship	In Person (United States)	September 2026
RevenueCat Shipaton 2026	RevenueCat	Global Hackathon	Online (Global)	October 1
Kaggriculture AI Agent Competition	Kaggle	AI Competition	Online (Global)	September 30

Resource Highlight

Resource	Why Check It Out
IBM SkillsBuild	A free online learning platform offering courses and learning pathways across AI, cybersecurity, data, cloud, and other technology skills, with learning available in 20+ languages.

Community Finds

Shared By	Find	Type	Location / Format
Konark Sharma (@konark_13)	The Agent Harness Hackathon	AI Hackathon	Online (Global) / In Person (San Francisco)

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from previous editions that are still accepting applications.

I've already covered these in detail, so I won't repeat everything here. If any of them catch your attention, you can find the full overview, eligibility details, and application links in the original edition.

Opportunity	Organization	Type	Format	Deadline	Featured In
FR8	FR8	Builder Residency	In Person	Rolling	Edition #2
Anthropic Fellows Program	Anthropic	AI Research Fellowship	In Person	Rolling	Edition #4
Sentient Open Source AGI Grant Program	Sentient Foundation	AI Grant & Investment	Remote	Rolling	Edition #6
Solo Grants	Solo Grants	Microgrant for Solo Builders	Remote	Rolling	Edition #7
Claude for Open Source	Anthropic	Open Source Program	Remote	Rolling	Edition #7
The Bridge	Entrepreneurs First	Founder Residency	In Person	August 30	Edition #9
Develop for Good Winter 2027	Develop for Good	Student Volunteer Program	Remote	August 31	Edition #10
Claude Community Ambassadors	Anthropic	Community Ambassador Program	Hybrid (Local Events)	Rolling	Edition #11
Z Fellows	Z Fellows	Founder Fellowship	Hybrid	Rolling	Edition #11
Magnificent Grants	Magnificent Grants	Grant & Fellowship	Hybrid	Rolling	Edition #12
AI Societal Impact Lab Autumn 2026 Fellowship	AI Societal Impact Lab	Fellowship	Remote	September 4	Edition #12
CodePath Fall 2026 Courses	CodePath	Technical Education Program	Virtual	August 23	Edition #12

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 a16z Alpha Fellowship

Who it's for: Technical students and recent graduates, generally within three years of graduating, who want to work at a fast-growing startup or try building their own company.

What stands out: The Alpha Fellowship gives you two options. You can join a fast-growing a16z portfolio company as a full-time software engineer, or you can apply through the Founder Track and get $20,000 in equity-free funding to start building your own company. The Founder Track can also lead to up to $250,000 in additional investment.

I wanted to include this because it's a pretty unusual opportunity for someone early in their career. You're not just getting another fellowship or a short course. You can either get a full-time engineering role at a growing startup or get funding to try building something yourself.

Another thing I like is that you don't need to already have a company or a complete team for the Founder Track. You apply as an individual, and if you already know who you want to build with, each person can apply separately and mention the others.

The fellowship itself is also very hands-on. It runs for eight weeks in person and includes a kickoff retreat, founder AMAs, small-group dinners, and access to the a16z speedrun community and events.

One important thing to know is that this is not globally accessible. You need valid U.S. authorization to work full time, and accepted fellows need to relocate for the in-person fellowship. I know I usually try to share opportunities that are accessible to people from different parts of the world, but I still wanted to include this one because the opportunity itself is too interesting to leave out. If you're eligible, definitely take a look.

Startup Track: Full-time software engineering role with salary and equity

Founder Track: $20,000 equity-free grant + up to $250,000 in additional investment

Fellowship: 8 weeks, in person

Who can apply: Technical students and recent graduates, generally within 3 years of graduation

Important: U.S. full-time work authorization is required.

Applications: Cohort 2 applications open in September 2026. Early applicants can submit now for priority consideration.

💡 Tip: You can apply to both tracks. If you're already building with a team, each person should apply separately and mention the other team members.

🔗 Learn More

📌 RevenueCat Shipaton 2026

Who it's for: Developers, students, indie builders, and teams who want to build and launch a mobile app.

What stands out: The RevenueCat Shipaton 2026 is a global online hackathon where the goal isn't just to build something. You actually have to launch a new app during the competition and use RevenueCat for monetization.

And the prize pool is pretty big. There is more than $740,000 in cash prizes, along with things like a Times Square billboard, media coverage, and a chance to attend RevenueCat's App Growth Annual conference in New York.

I wanted to include this because I like hackathons where you actually have to ship something people can use. Instead of building a prototype and stopping there, Shipaton gives you a reason to take the idea all the way to launch.

Another thing I like is that there are a lot of different categories, so you don't necessarily have to compete only for the main prize. There are awards for things like design, gaming, social good, growth, building in public, monetization, and a student-only Next Gen Award.

There are also sponsor prizes from companies like JetBrains, OneSignal, Replit, Samsung, Stripe, and others.

Prize Pool: $740,000+ | Format: Online, global hackathon

Build Window: August 1 – September 30, 2026 | Submission Deadline: October 1, 2026

Team: Individual or team participation

Who can participate: Ages 13–99, with some countries and territories excluded.

Main Requirement: Build and publicly launch a new app during the Shipaton window and integrate RevenueCat for monetization.

🔗 Learn More & Join the Hackathon

📌 Kaggriculture AI Agent Competition

Who it's for: Developers, students, AI engineers, machine learning practitioners, and anyone interested in experimenting with AI agents. Some Python experience will be helpful.

What stands out: Kaggriculture is a Kaggle competition where you build an AI agent that manages a virtual farm and competes against other agents on a live leaderboard.

Your agent has to make decisions over hundreds of turns, balancing things like farming, trading, expansion, resources, and changing market conditions to try to make the most profit.

I wanted to include this because it's very different from the Kaggle competitions you usually see. You're not just given a dataset and asked to build the best prediction model. You're actually building an agent that has to make decisions, plan ahead, and react to what is happening around it.

Another thing I like is that it gives you a way to experiment with agent design without needing to build a huge project from scratch. You're working inside a game-like environment, but you're still thinking about things like strategy, optimization, resource management, and planning.

And honestly, the farming part just makes it more fun than another competition where you're staring at a dataset all day.

Prize Pool: $50,000 USD ($5,000 each for the top 10 teams)

Format: Online Kaggle competition

Competition Start: July 29, 2026 | Entry Deadline: September 23, 2026

Final Submission Deadline: September 30, 2026

Team Size: Individual or team

🔗 Learn More & Join

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

IBM SkillsBuild

Who it's for: Students, recent graduates, and anyone who wants to build skills in areas like AI, cybersecurity, data, cloud, and other emerging technologies.

What stands out: IBM SkillsBuild is a free online learning platform with courses, learning pathways, hands-on learning, and digital credentials.

I wanted to include this because it's one of those resources that's worth bookmarking. If you're trying to learn something new alongside school, work, or a job search, there are a lot of different topics you can explore without paying for another course.

You can find content on things like generative AI, AI ethics, cybersecurity, data, cloud computing, and quantum computing, so there's a good amount to explore depending on what you're interested in.

Another thing I like is that it isn't just video courses. There are also hands-on labs, learning pathways, digital credentials, and virtual events.

Cost: Free | Format: Online | Languages: 20+ languages

Topics: AI, Generative AI, Cybersecurity, Data, Cloud, Quantum Computing, and more

Credentials: Digital credentials are available for selected activities and pathways.

🔗 Explore IBM SkillsBuild

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

The Agent Harness Hackathon

Shared by Konark Sharma (@konark_13).

Konark shared the Agent Harness Hackathon, a seven-day online hackathon from WeMakeDevs in collaboration with TrueFoundry.

The challenge is to build an AI agent using TrueForge, TrueFoundry's open-source agent harness. The idea is to build an agent that can actually take actions, connect to real tools, safely run generated code, and ask for human approval before doing something irreversible.

I wanted to include this because I like that it's focused on building agents that can actually do things, rather than just answering questions. You can build something like a code review agent, incident responder, research assistant, analytics agent, or come up with your own idea.

There are several categories to compete in, including Best Use of TrueForge, Best Code Quality, and Best UI. There are also prizes for the best blog and social posts.

The total prize pool is $10,000, including an NVIDIA DGX Spark, Mac Mini, iPad, and other prizes.

Format: Online, with an optional in-person event in San Francisco

Dates: August 24–30, 2026

Team Size: Solo or teams of up to 4

Prize Pool: $10,000

Submission Deadline: August 30, 2026 at 8:00 PM London time

Who can participate: Developers, students, builders, and anyone interested in AI agents.

🔗 Learn More & Register

Thank you to Konark Sharma (@konark_13) for thinking of the radar and sharing this with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

Transparency note: I used AI (Gemini) to create the banner image for this article.

I also used AI to help with grammar and wording. English isn't my first language, and sometimes I know exactly what I want to say but struggle to find the right words. I used AI to help me express those thoughts more clearly, especially when explaining opportunities and making sure the information was easy to follow.

The research, ideas, experiences, perspectives, and decisions about what to include are my own. As always, I recommend checking the official links before applying, since deadlines and program details can change.

Dev Opportunity Radar #12: $10K+ Magnificent Grants, Free CodePath Courses, AI Societal Impact Lab Fellowship, and Interview Prep 101

Hemapriya Kanagala — Fri, 14 Aug 2026 17:48:56 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features Magnificent Grants, the AI Societal Impact Lab Autumn 2026 Fellowship, CodePath's Fall 2026 Courses, and Interview Prep 101, a free workshop recording from CodePath's Emerging Engineers Summit packed with practical advice for technical interviews.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features an opportunity shared by Konark Sharma (@konark_13). Thank you, Konark, for recommending the Zero Downtime Hackathon. I always enjoy seeing readers help others discover opportunities they might otherwise have missed, and I hope this section continues to grow.

If you've discovered something through the radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Magnificent Grants
- 📌 AI Societal Impact Lab Autumn 2026 Fellowship
- 📌 CodePath Fall 2026 Courses
📚 Resources Worth Checking Out
- CodePath: Interview Prep 101
🌟 Community Finds
- Zero Downtime Hackathon
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Format	Deadline
Magnificent Grants	Magnificent Grants	Grant & Fellowship	Hybrid	Rolling
AI Societal Impact Lab Autumn 2026 Fellowship	AI Societal Impact Lab	Fellowship	Remote	September 4
CodePath Fall 2026 Courses	CodePath	Technical Education Program	Virtual	August 23

Resource Highlight

Resource	Why Check It Out
CodePath: Interview Prep 101	A free workshop recording covering practical interview strategies, communication, structuring answers, and techniques to help you perform more confidently in technical interviews.

Community Finds

Shared By	Find	Type	Format
Konark Sharma (@konark_13)	Zero Downtime Hackathon	AI Hackathon	In Person (San Francisco)

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from previous editions that are still accepting applications.

Opportunity	Organization	Type	Format	Deadline	Featured In
FR8	FR8	Builder Residency	In Person	Rolling	Edition #2
Anthropic Fellows Program	Anthropic	AI Research Fellowship	In Person	Rolling	Edition #4
Sentient Open Source AGI Grant Program	Sentient Foundation	AI Grant & Investment	Remote	Rolling	Edition #6
Solo Grants	Solo Grants	Microgrant for Solo Builders	Remote	Rolling	Edition #7
Claude for Open Source	Anthropic	Open Source Program	Remote	Rolling	Edition #7
AIAF Fellowship 2026	AIAF × AE Studio	AI Research Fellowship	Remote	August 17	Edition #8
The Bridge	Entrepreneurs First	Founder Residency	In Person	August 30	Edition #9
Develop for Good Winter 2027	Develop for Good	Student Volunteer Program	Remote	August 31	Edition #10
Claude Community Ambassadors	Anthropic	Community Ambassador Program	Hybrid (Local Events)	Rolling	Edition #11
Z Fellows	Z Fellows	Founder Fellowship	Hybrid	Rolling	Edition #11

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Magnificent Grants

Who it's for: Builders, founders, researchers, creators, and ambitious young people aged 25 or younger who are working on ambitious ideas, side projects, startups, or other high-impact work. You don't need to have a company or drop out of school to apply, although showing meaningful progress on what you're building will strengthen your application.

What stands out: Magnificent Grants awards $10,000+ grants to a small group of young builders each year, along with two years of mentorship from experienced founders and investors. Unlike many startup programs, the grant comes with no equity requirement, allowing recipients to continue building independently while receiving ongoing support.

I wanted to include this because it's refreshing to see a program that's willing to support people at a very early stage. Many fellowships expect applicants to already have traction, funding, or a well-established company. Here, the focus seems to be on identifying exceptional people with ambitious ideas and helping them move faster.

Another thing I like is that the support extends well beyond the grant itself. Fellows become part of a two-year community where they receive mentorship, meet other grantees, and continue building relationships with experienced founders and investors. The fact that you don't need to drop out of school or leave your job also makes it accessible to more people who are still exploring their ideas.

Grant: $10,000+ (no equity required)

Program Duration: Two years

Who can apply: Anyone aged 25 or younger from anywhere in the world.

Application Deadline: Rolling

🔗 Learn More | 🔗 Apply

📌 AI Societal Impact Lab Autumn 2026 Fellowship

Who it's for: Recent graduates, early-career researchers, and people interested in exploring how AI is shaping society. While the fellowship is primarily aimed at applicants from the social sciences and humanities, people from STEM, the arts, and other backgrounds are also encouraged to apply. Previous AI experience is welcome but not required.

What stands out: The AI Societal Impact Lab Autumn 2026 Fellowship is a remote 8 to 10-week program that brings together people interested in the societal and political implications of AI. Fellows participate in weekly discussions, receive mentorship, work on a collaborative group project, attend workshops and seminars, and write their own analytical piece that will be published on the AI Societal Impact Lab website.

I wanted to include this because most AI fellowships focus on building models or developing technical skills. This one takes a different approach by exploring the broader questions around how AI is changing society and encouraging thoughtful discussions from people with different backgrounds.

Another thing I like is that you don't need to be an AI expert to apply. The fellowship welcomes people from a variety of disciplines and creates a space where different perspectives can come together. If you've been curious about AI governance, ethics, public policy, or the social impact of emerging technologies, this looks like a great place to learn and contribute.

Program Duration: Approximately 8 to 10 weeks (starting in mid to late September 2026)

Time Commitment: Around 4 hours per week

Format: Fully remote

Benefits: Mentorship, workshops and seminars, collaborative group project, publication on the AI Societal Impact Lab website, and a community of fellows.

Who can apply: Recent graduates, early-career researchers, and applicants from a wide range of backgrounds, including humanities, social sciences, STEM, and the arts.

Application Deadline: September 4, 2026 (rolling until filled)

🔗 Learn More | 🔗 Apply

📌 CodePath Fall 2026 Courses

Who it's for: Students and recent graduates in computer science or software-related programs who want to strengthen their technical skills, prepare for internships and full-time roles, and learn in a structured, instructor-led environment. Some pathways have additional prerequisites, so be sure to review the requirements before applying.

What stands out: CodePath is offering its Fall 2026 cohort of free, instructor-led technical courses in Applied AI Engineering, Technical Interview Prep, Cybersecurity, and Web Development. Each pathway runs virtually for around 10 weeks, combining live classes, hands-on assignments, projects, and mentorship from engineers at leading technology companies.

I wanted to include this because I've personally completed two CodePath courses: Intermediate Cybersecurity (with Honors) and Intro to Technical Interview Prep. They were genuinely some of the most valuable learning experiences I've had. The courses aren't passive video lectures. Every week includes live, interactive sessions, breakout discussions with peers and teaching assistants, and practical assignments that push you to apply what you've learned.

One thing I also appreciate is that CodePath expects commitment. Getting accepted isn't the final step. Every applicant is required to complete pre-work as part of the admissions process, and it takes real time and effort. It's one of the reasons the community stays so engaged. Everyone begins the course having already demonstrated they're willing to put in the work.

Another thing I like is that the experience continues well beyond the classroom. After completing a course, students gain access to career services, a large network of students and alumni, industry mentors, recruiting events, career fairs, workshops with leading technology companies, and an active Slack community where people continue learning and helping one another.

Available Pathways:

Applied AI Engineering
Technical Interview Prep
Cybersecurity
Web Development

Duration: Approximately 10 weeks

Format: Virtual, instructor-led

Cost: Free

Who can apply: Students aged 18+ who will be enrolled in a U.S. college or university during Fall 2026 in a computer science or software-related degree program. WEB101 and CYB101 are open to students from all majors. Recent computer science graduates are also eligible. Individual courses may have additional prerequisites.

Application Deadline: August 23, 2026

💡 Tip: Don't wait until the last minute. After submitting your application, you'll need to complete the required pre-work before your application can be reviewed. The pre-work takes time, so applying early gives you the best chance of completing everything before the deadline.

🔗 Learn More & Apply

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

CodePath: Interview Prep 101 (Free Workshop Recording)

Who it's for: Students, new graduates, internship applicants, and anyone preparing for software engineering interviews.

What stands out: This free session from CodePath's 2024 Emerging Engineers Summit focuses on the parts of interviewing that people often overlook. Instead of diving straight into coding problems, it covers how to make a strong first impression, communicate your thought process, structure your answers, and approach technical interviews with more confidence.

I wanted to include this because interview preparation isn't just about solving LeetCode problems. Being able to clearly explain your thinking, communicate with your interviewer, and approach questions in a structured way can make a huge difference. I found this session practical and easy to follow, especially for people preparing for internships or new graduate roles.

Another thing I like is that it's only around an hour long, making it an easy resource to watch before interview season. If you're already practicing coding problems, this is a nice complement that focuses on the communication side of technical interviews.

Format: Free recorded workshop

Level: Beginner to Intermediate

Topics Covered: Technical interview preparation, communication, structuring answers, interview tips, and best practices.

🔗 Watch the Session

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

Zero Downtime Hackathon

Shared by Konark Sharma (@konark_13).

Konark shared the Zero Downtime Hackathon, and I thought it would be worth passing along for anyone interested in building AI applications.

The hackathon focuses on building resilient AI systems rather than just prototypes. Participants will use Port for project organization, Bright Data Scraper Studio for web data extraction, and SigNoz for observability to create an end-to-end application, pipeline, or AI agent. Projects will also be judged by engineers from Anthropic.

I wanted to feature this because it encourages participants to think beyond simply getting an AI application to work. It emphasizes planning, monitoring, and building systems that can continue running even when things change, which are skills that become increasingly important as AI projects move into production.

There are also some exciting prizes, including an NVIDIA DGX Spark AI Supercomputer for the overall winner, along with track prizes for the best integrations with each sponsor tool.

Who it's for: Developers, students, builders, and AI enthusiasts interested in building production-ready AI applications.

Format: In-person hackathon (San Francisco, California)

Event Date: August 22, 2026

Team Size: 1 to 4 participants

🔗 Learn More & Register

Thank you to Konark Sharma (@konark_13) for thinking of the radar and sharing this with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post

What Would You Tell Someone Early in Their Career?

Hemapriya Kanagala — Mon, 10 Aug 2026 12:42:18 +0000

📌 TL;DR

I'm still early in my own career, and lately I've been thinking about how much advice we hear when we're starting out.

Learn this. Build that. Get certified. Network more. Specialize. Don't specialize too early. Work harder. Protect your work-life balance.

Sometimes it feels like everyone has an answer 😄

But careers are long, and I think some lessons only really make sense after you've lived through them.

That's what made me think about this community.

DEV has people at so many different stages of their careers, with completely different experiences. Some are just starting. Some have been working for a few years. Others have decades of experience. Some have changed careers entirely, managed teams, built companies, made mistakes, started over, burned out, found work they love, or discovered that what they thought they wanted wasn't actually right for them.

That's a lot of experience in one place.

So this post is a little different. I'll share some of what I'm learning while being early in my own career, but mostly, I want to listen to you.

What would you tell someone early in their career?

And there's no limit here 😄 If one thing comes to mind, share one. If you have five, ten, or a whole list of things experience has taught you, I'd genuinely love to hear them all. They don't have to be about the same topic either.

Talk about whatever you feel is worth passing on. It can be technical, personal, or somewhere in between: work, money, learning, managers, communication, confidence, health, work-life balance, mistakes, opportunities, relationships, or something completely different.

And if you're early in your career too, you're absolutely part of this conversation. Share something you've already learned, something you wish you'd known when you started, or something you're still trying to figure out.

And if you have a question, ask it. There might be someone here who's already been through exactly what you're wondering about and can share what they learned. And if it's something I happen to know or have experienced too, I'll be happy to share what I can.

My hope is that we can turn the comments into a place where people at different stages of their careers can ask questions, share experiences, and learn from each other 💙

📚 Table of Contents

This Post Is a Little Different
I'm Learning That There Probably Isn't One Correct Way to Build a Career
There's So Much About a Career That Doesn't Fit on a Resume
Some of You Have Already Lived the Years the Rest of Us Are Just Beginning
So... What Would You Tell Someone Early in Their Career?
It Doesn't Have to Be Career Advice in the Traditional Sense
Why I'm Asking the Community Instead of Writing an Advice Post
I Have a Feeling the Best Part of This Post Might Be the Comments

This Post Is a Little Different

I had a slightly different idea for this post.

Usually when I sit down to write something, I have a topic I want to explore, an experience I want to share, or, apparently lately, a technical report that sends me down a rabbit hole for several hours 😄

But this time, I want to write a little less and listen a little more.

I'm still early in my own career. There are things I understand much better today than I did a few years ago, but I'm also pretty sure there are things I'll look back on five or ten years from now and think, “I really wish I had understood that earlier.”

Maybe someone has already told me some of those things and I haven't had enough experience yet to understand why they matter. Or maybe they're lessons I'll only learn by going through something myself.

Either way, it made me think about something I really like about communities like DEV: we're all at completely different points in our careers.

Some people here are just figuring out what they want to do, while others have been working for years or even decades. And even within that, no two paths look exactly the same. People change fields, become managers, go back to being individual contributors, build companies, take career breaks, start over, or watch their carefully planned career go somewhere completely unexpected.

Every one of those paths comes with lessons.

And that's what made me wonder:

What could those of us who are early in our careers learn from everyone else's experiences?

I'm Learning That There Probably Isn't One Correct Way to Build a Career

When you're early in your career, it's surprisingly easy to feel like you're doing something wrong.

You look around and someone seems to be learning faster than you, someone else just got promoted, and another person landed an opportunity you'd love to have. Some people already seem to know exactly what they want to specialize in, while others have incredible portfolios or somehow sound completely confident about where their careers are going.

And meanwhile, you might still be sitting there thinking, “I don't even know exactly what I want to be doing three years from now” 😄

I've definitely spent time worrying about things like that myself. But I'm slowly realizing that careers probably aren't as linear as they sometimes look from the outside.

People change roles, discover that something they thought they'd love isn't actually for them, or become interested in something they never planned to pursue. They take breaks, make mistakes, change their minds, and sometimes start again. An opportunity that looks perfect might teach you what you don't want, while something that feels like a complete detour might end up being useful years later in a way you couldn't possibly have predicted at the time.

I think that's part of what makes comparing career paths so difficult. We usually see where someone is now, but we don't necessarily see all the changes, uncertainty, mistakes, and unexpected turns that happened along the way.

I'm still figuring out my own path, so I'm definitely not going to tell anyone exactly how they should build theirs. But one thing I'm trying to become more comfortable with is not having the entire map yet.

Maybe knowing where I want to go next is enough for now.

There's So Much About a Career That Doesn't Fit on a Resume

When we're starting out, a lot of our attention naturally goes toward the things we can show: the skills we're learning, projects we're building, certifications we're earning, our education, experience, and eventually the achievements we can point to.

And of course those things matter.

But the longer I think about careers, the more I wonder whether some of the lessons that shape them the most are things we'll never put on a resume.

Things like knowing when to ask for help instead of spending another five hours trying to solve something alone. Learning how to disagree with someone without turning that disagreement into a fight. Being able to say “I don't know” without immediately worrying that everyone will suddenly think you don't belong there.

Then there are decisions that don't always have an obvious right answer. How do you know whether an opportunity is genuinely good for you or just looks impressive from the outside? When should you negotiate, and when should you walk away? How do you deal with a bad manager, or eventually learn to become a good one yourself?

And some of the questions go beyond work itself. How much of your identity should your career take up? How do you stay ambitious without feeling like every free hour needs to become another opportunity to be productive? How do you take care of your health while still trying to build the future you want?

I don't think there's one course we can take that neatly teaches all of that.

A lot of it seems to come from living through things, making choices, getting some of them wrong, watching other people go through their own experiences, and slowly figuring out what matters to us along the way.

And that's exactly why I wanted to write this post.

Some of You Have Already Lived the Years the Rest of Us Are Just Beginning

That's something I think is pretty amazing when you really stop and think about it.

Someone reading this might be preparing for their very first interview, while someone else might be three years into their career. Another person might have fifteen or twenty years of experience, and someone else might have spent decades doing one thing only to eventually decide they wanted to start over in a completely different direction.

And all of that experience means much more than knowing different technologies or having different job titles.

People have worked with managers they loved and managers who taught them exactly what they wouldn't want to do if they ever became one themselves. They've made career decisions that worked out beautifully and others they probably wouldn't make again. They've negotiated salaries, accepted opportunities, turned others down, stayed somewhere too long, left somewhere too early, found mentors, and maybe eventually became mentors for someone else.

They've also had time to discover which worries actually mattered. Some of the things that felt incredibly important at the beginning of their careers might barely matter to them now, while things they hardly thought about back then may have turned out to matter much more than they expected.

And that's the kind of knowledge I'm curious about.

I'm not looking for another perfect career roadmap or a formula that says, “Do these seven things and you'll be successful.”

I'm much more interested in what people learned from actually going through it: the good decisions, the mistakes, the unexpected turns, and the things they understand now that they couldn't have understood when they were just beginning.

So... What Would You Tell Someone Early in Their Career?

This is the part where I hand the post over to you 😄

If you've been working for a while, I'd love for you to think back to the beginning of your career and everything you've experienced since then.

What do you know now that you wish you'd understood earlier? And what would you want someone early in their career to know?

Please don't feel like you have to choose just one thing either. You might have something to say about technical skills and something completely different to say about work-life balance. Your lessons might be about managers, communication, money, confidence, networking, changing jobs, burnout, asking for help, choosing opportunities, dealing with mistakes, knowing when to walk away, or something I haven't even thought to mention.

Maybe there's something you spent far too much time worrying about early on that you now realize wasn't nearly as important as it seemed. Or perhaps there's something you barely thought about at the beginning that you now wish you'd taken much more seriously.

And if your advice completely contradicts someone else's advice here, that's okay too. I'd actually love to see those different perspectives. What worked beautifully for one person might have been completely wrong for someone else.

You also don't need twenty years of experience to contribute. If you're early in your career too, you're just as much a part of this conversation. You might already have learned something from your first job, an opportunity you took, a mistake you made, a difficult experience, or simply something that's helped you along the way.

And if you feel like you're still figuring everything out, you don't need to have advice at all.

You can ask instead.

What do you wish you could ask someone further along in their career?

There might be someone reading this who remembers having exactly the same question and has already lived through the experience you're wondering about.

That's really what I'm hoping this discussion can become: not experienced people talking at people who are just beginning, but people at different stages of their careers learning from each other.

It Doesn't Have to Be Career Advice in the Traditional Sense

When we hear “career advice,” it's easy for the conversation to immediately become about what skill to learn next, which certification to get, or which technology to master.

And those conversations can absolutely be useful. But I think a career is much bigger than the technical skills required to do a job.

Maybe experience taught you to protect your weekends, start saving money earlier, ask more questions, find mentors, document your work, take your health seriously, make sure you have a life outside work, or stop comparing your entire career with the small part of someone else's life you see online.

Maybe it's something completely unexpected that I haven't even thought to mention.

Actually, those unexpected answers might be the ones I'm most curious to read.

Because sometimes the advice that stays with us isn't the advice we went looking for in the first place.

Why I'm Asking the Community Instead of Writing an Advice Post

I could have written an article called “Things You Should Know Early in Your Career.”

But that wouldn't feel quite right coming from me.

I'm early in my own career too. I have things I've learned so far and experiences that have changed how I think, but I also know there are situations I simply haven't lived through yet.

Instead of pretending otherwise, I'd rather ask people who have lived through them.

And experience doesn't only mean years worked.

Someone with twenty years of experience might share something that helps someone with two. Someone with two years might remember the uncertainty of starting out much more clearly. And someone entering the workforce today is experiencing things around AI, remote work, learning, and hiring that didn't exist in quite the same way ten or twenty years ago.

Different stages of a career teach us different things, and that's what makes this much more interesting to me than trying to write a career guide myself.

I Have a Feeling the Best Part of This Post Might Be the Comments

And honestly, that's what I'm hoping for 😄

I've had conversations here where someone left a comment that made me look at something completely differently, introduced me to an idea I hadn't considered, or brought years of experience to something I had only recently started thinking about.

That's one of my favorite things about writing in a community like DEV. You don't just publish something and walk away. Sometimes what you write starts a conversation, and that conversation ends up teaching you just as much as the article itself.

So maybe this post can become a little collection of experiences.

Not a definitive career guide.

Not a roadmap everyone is supposed to follow.

Just people at different stages of their lives and careers sharing what they've learned along the way.

Someone might come across a comment and think, “I really needed to hear that.” Someone else might discover an idea they hadn't considered before, or read a perspective that stays with them until they encounter that situation themselves.

And I fully expect to be taking notes too 😄

So I'll leave this one with all of you:

What would you tell someone early in their career?

Share as much or as little as you'd like. If you have one lesson, I'd love to hear it. If years of experience have given you five or ten things you wish you'd known earlier, please share those too.

And if you're early in your career too, I'd love to hear from you just as much. Tell us what you've already learned, what you're struggling with, or what you wish you could ask someone who's been where you are now.

There isn't one correct way to build a career, and I don't expect there to be one correct answer here either.

In fact, I think the different answers are what could make this discussion useful.

So, what has your career taught you that you'd pass on to someone who's still near the beginning? I'd really love to hear it 💙

Transparency note: The banner image for this article was generated using DEV Community's built-in AI image generation feature. I wrote the prompt, generated multiple variations, and selected the final image used in this post.

I also used AI to help check my grammar and make sure I was expressing my thoughts clearly. English isn't my first language, so sometimes I know what I want to say but need a little help making sure the wording comes across correctly. The ideas, experiences, and perspectives in this article are my own.

Dev Opportunity Radar #11: Claude Community Ambassadors, YC Startup Internship Expo, Z Fellows, and Stanford's AI Agents Course

Hemapriya Kanagala — Fri, 07 Aug 2026 15:52:30 +0000

TL;DR

Welcome back to Dev Opportunity Radar.

This is a weekly series where I share opportunities, resources, communities, and interesting finds that I come across, with the goal of helping people discover things they might otherwise miss.

This week's edition features the Claude Community Ambassadors Program, the YC Startup Internship Expo, Z Fellows, and CS329A: Self-Improving AI Agents, a free graduate-level Stanford course exploring the latest research behind modern AI agents.

If you're new to the series, you can also browse previous editions, search past opportunities, and explore Community Finds, Reader Updates, and Resources Worth Checking Out on the Dev Opportunity Radar website. I've also written a short post about why I built it. You'll find links to both at the end of this article.

This week's 🌟 Community Finds section features opportunities shared by Francis (@francistrdev) and Konark Sharma (@konark_13). I always love seeing readers help others discover opportunities they might otherwise have missed.

If you've discovered something through the radar, I'd love to hear about it. Whether you applied to an opportunity, attended an event, joined a community, completed a program, built something, or found a resource you hadn't seen before, I'd be happy to feature your experience in a future 💙 Reader Updates section (with your permission).

And if you've come across an opportunity, resource, community, program, event, or anything else you think deserves more attention, feel free to share it in the comments.

If I feature one of your 🌟 Community Finds in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

⚡ Quick Scan
🔄 Still Open From Previous Editions
📍 This Week's Opportunities
- 📌 Claude Community Ambassadors
- 📌 YC Startup Internship Expo
- 📌 Z Fellows
📚 Resources Worth Checking Out
- CS329A: Self-Improving AI Agents
🌟 Community Finds
- CALEC Volunteer and Internship Opportunities
- JPMorganChase Code for Good Singapore & Software Engineer Program
💙 Reader Updates
👋 Until Next Friday
🌐 Dev Opportunity Radar Website

⚡ Quick Scan

Opportunities

Opportunity	Organization	Type	Format	Deadline
Claude Community Ambassadors	Anthropic	Community Ambassador Program	Hybrid (Local Events)	Rolling
YC Startup Internship Expo	Y Combinator	Startup Career Expo	In Person (San Francisco)	August 15 (Event)
Z Fellows	Z Fellows	Founder Fellowship	Hybrid	Rolling

Resource Highlight

Resource	Why Check It Out
CS329A: Self-Improving AI Agents	A free graduate-level Stanford course covering modern AI agent research, including constitutional AI, tool use, planning, reinforcement learning, memory, and evaluation frameworks.

Community Finds

Shared By	Find	Type	Format
Francis (@francistrdev)	CALEC Volunteer & Internship Opportunities	Volunteer & Internship Opportunities	Remote
Konark Sharma (@konark_13)	JPMorganChase Code for Good Singapore & Software Engineer Program	Hackathon & Internship	In Person (Singapore)

📝 A quick note: I spend a lot of time researching and verifying every opportunity before featuring it in Dev Opportunity Radar. However, deadlines, eligibility, program details, and application requirements can change after publication. Before applying, please take a few minutes to visit the official program page, review the latest information, and confirm that you're eligible.

🔄 Still Open From Previous Editions

Before we get into this week's opportunities, here are a few from previous editions that are still accepting applications.

Opportunity	Organization	Type	Format	Deadline	Featured In
FR8	FR8	Builder Residency	In Person	Rolling	Edition #2
Anthropic Fellows Program	Anthropic	AI Research Fellowship	In Person	Rolling	Edition #4
Sentient Open Source AGI Grant Program	Sentient Foundation	AI Grant & Investment	Remote	Rolling	Edition #6
Solo Grants	Solo Grants	Microgrant for Solo Builders	Remote	Rolling	Edition #7
Claude for Open Source	Anthropic	Open Source Program	Remote	Rolling	Edition #7
AIAF Fellowship 2026	AIAF × AE Studio	AI Research Fellowship	Remote	August 17	Edition #8
The Bridge	Entrepreneurs First	Founder Residency	In Person	August 30	Edition #9
Develop for Good Winter 2027	Develop for Good	Student Volunteer Program	Remote	August 31	Edition #10
OpenAI Student Collective	OpenAI	Campus Leadership Program	In Person (Campus-Based)	August 10	Edition #10

📍 This Week's Opportunities

Here are a few opportunities I came across this week that I thought were worth sharing.

📌 Claude Community Ambassadors

Who it's for: Developers, community organizers, meetup hosts, technical educators, and active Claude users who enjoy bringing people together and helping others learn. You don't need to be a professional developer, but you should have meaningful experience using Claude and a track record of community involvement.

What stands out: The Claude Community Ambassadors program gives community builders the opportunity to grow local Claude communities with support from Anthropic. Ambassadors can host meetups, workshops, hackathons, and other community events while receiving event funding, monthly API credits, ready-to-use event materials, swag, and promotion through Anthropic's channels.

I wanted to include this because most ambassador programs are focused on students. This one is different. It's designed for people who are already active in their local tech communities or online communities and want to help others explore Claude while building a stronger local developer ecosystem.

Another thing I like is that the program goes beyond hosting events. Ambassadors get early access to selected features, opportunities to participate in Builders Council sessions, and a direct way to share feedback from their communities with Anthropic's product teams. If you enjoy teaching, organizing events, or building developer communities, this could be a great opportunity to make an impact while connecting with other community leaders around the world.

Program Duration: Ongoing

Cost: Free

Benefits: Event sponsorship, monthly Claude API credits, exclusive swag, access to a private Ambassador Slack, pre-release features, and opportunities to provide product feedback.

Who can apply: Community builders and experienced Claude users from anywhere in the world. Professional developer experience is not required.

Application Deadline: Rolling

🔗 Learn More | 🔗 Apply

📌 YC Startup Internship Expo

Who it's for: Students looking for internships, co-ops, semester internships, or new graduate roles who will be in the San Francisco Bay Area on August 15. YC is also encouraging applications from exceptional high school builders.

What stands out: The YC Startup Internship Expo is a one-day, acceptance-only event where students can meet founders, engineers, and hiring teams from 50+ Y Combinator startups. Rather than browsing job boards, attendees get the chance to explore what it's actually like to work at early-stage AI startups, participate in coding and AI challenges, network with startup teams, and even interview onsite with more than 40 YC companies.

I wanted to include this because opportunities like this don't come around very often. Instead of applying online and hoping someone sees your resume, you get to meet founders and engineers face-to-face, ask questions, and learn how some of today's fastest-growing AI startups actually build products.

Another thing I like is that it's more than just a career fair. Alongside the startup expo, there will be coding and AI challenges, giveaways, food trucks, and onsite interviews happening throughout the day. If you're already spending the summer in San Francisco or studying in the Bay Area, this could be a great opportunity to explore internships while connecting with people building at the forefront of AI.

Event Date: August 15, 2026

Location: San Francisco, California

Cost: Free (application required)

Who can apply: Students interning in San Francisco, students attending Bay Area colleges or universities, and exceptional high school builders who will be in the Bay Area on the event date.

Application Process: Acceptance only

Travel Support: No travel sponsorship is provided.

🔗 Learn More & Apply

📌 Z Fellows

Who it's for: Technical builders, aspiring founders, students, recent graduates, researchers, engineers, and anyone working on a side project or startup. You can apply solo or with a team, and you don't need to have a company yet.

What stands out: Z Fellows is a one-week founder program that brings together a small group of ambitious builders with founders and investors behind companies like Figma, Vercel, Instacart, Netflix, DoorDash, and AngelList. The week is designed to help early builders learn from experienced founders, connect with other ambitious people, and accelerate whatever they're working on. The support doesn't stop after the program. Z Fellows continues helping founders throughout their journey.

I wanted to include this because most startup programs expect you to already have a company, traction, or funding. Z Fellows takes a different approach. It's built for people who are still early and simply have the ambition to build something meaningful. Whether you're working on a side project, exploring an idea, or already building a startup, the focus is on backing people with potential before everyone else notices them.

Another thing I like is how flexible the program is. You don't need to quit your job, leave school, or even have a co-founder to apply. The program is mostly virtual, with the final day taking place in San Francisco or New York City, making it accessible to builders from around the world. Applicants can also choose to participate without funding or accept an optional $10,000 investment through a SAFE agreement if it makes sense for their company.

Program Duration: One week (multiple cohorts each year)

Format: Mostly virtual, with the final day held in San Francisco or New York City

Funding: Optional $10,000 investment through a SAFE agreement

Who can apply: Builders from anywhere in the world. No age restrictions. You can apply as an individual or as a team, and you don't need a company to get started.

Application Deadline: Rolling

🔗 Learn More | 🔗 Apply

📚 Resources Worth Checking Out

Not every useful find comes with an application deadline.

Here's one resource worth checking out this week.

CS329A: Self-Improving AI Agents

Who it's for: Developers, AI engineers, researchers, graduate students, and anyone who already has a basic understanding of large language models and wants to dive deeper into modern AI agent research.

What stands out: CS329A: Self-Improving AI Agents is a graduate-level course from Stanford University that explores how AI agents can continuously improve through interaction with themselves and their environment. The lecture recordings cover many of the ideas shaping today's AI landscape, including constitutional AI, reinforcement learning, test-time scaling, tool use, memory, planning, multimodal agents, and evaluation frameworks for agentic systems.

I wanted to include this because there's a lot of discussion about AI agents right now, but it can be difficult to separate hype from the underlying concepts. This course is taught by researchers working in the field and provides a structured way to understand many of the techniques behind modern agent systems.

Another thing I like is that the course doesn't focus on a single framework or API. Instead, it explores the research ideas that are driving the next generation of AI agents, making it a resource that's likely to stay useful even as tools continue to evolve.

Format: Free online lecture recordings

Level: Graduate / Advanced

Topics Covered: Self-improving LLMs, Constitutional AI, reinforcement learning, test-time scaling, tool use, memory, planning, multimodal agents, and AI evaluation.

🔗 Watch the Lecture Playlist

🌟 Community Finds

One of my favorite things about this series has been seeing people share opportunities, communities, and resources that others might not have discovered otherwise.

CALEC Volunteer and Internship Opportunities

Back in Edition #6, I featured CALEC's remote volunteer and internship opportunities after Francis (@francistrdev) shared them with the community. If you missed it, you can find the full overview and all the details in Edition #6.

Since sharing this opportunity with the community, Francis has also published a dedicated DEV post about the Full-Stack Developer volunteer role at CALEC. If that role interests you, it's worth checking out for his perspective and some additional information about what the team is looking for.

🔗 Francis' post

🔗 CALEC Volunteer & Internship Opportunities | 🔗 Edition #6

JPMorganChase Code for Good Singapore & Software Engineer Program

Shared by Konark Sharma (@konark_13).

Konark shared two opportunities from JPMorganChase that I thought would be worth passing along for students interested in software engineering.

The first is Code for Good Singapore, a two-day social impact hackathon where participants work alongside JPMorganChase engineers to build technology solutions for nonprofit organizations. Besides the experience itself, the event also serves as a pathway to the company's 2027 Early Careers Software Engineer Program in Singapore.

He also shared the 2027 Software Engineer Program - Summer Analyst (Singapore). Interns work on real software engineering projects, collaborate with experienced engineers, receive mentorship and training, and may be considered for a full-time graduate role based on their performance.

I thought these were worth featuring because they offer more than just a hackathon or an internship. Together, they provide a chance to gain hands-on experience, connect with engineers, and potentially open the door to future opportunities at one of the world's largest technology employers.

Who it's for: Students pursuing a bachelor's degree in computer science or a related field who are interested in software engineering opportunities in Singapore.

Application Deadlines:

Code for Good Singapore: September 18, 2026
Software Engineer Program - Summer Analyst (Singapore): October 29, 2026

🔗 Code for Good Singapore

🔗 Software Engineer Program - Summer Analyst (Singapore)

Thank you to Francis (@francistrdev) and Konark Sharma (@konark_13) for thinking of the radar and sharing these with the community. It really means a lot 💙

I'd love for this section to keep growing.

If you've come across an opportunity, fellowship, grant, hackathon, conference, community, resource, or anything else you think more people should know about, feel free to share it in the comments.

If I feature it in a future edition, I'll always make sure to credit you. If you discovered it, that recognition belongs to you.

One small request: If you're sharing an opportunity, please avoid posting raw URLs directly in the comments. DEV sometimes filters them before I get a chance to see them.

A short description alongside the link makes it much easier for me to review and potentially feature it in a future edition.

💙 Reader Updates

I'm looking forward to this section gradually growing over time, and I'd still love to hear from you.

One of my favorite parts of writing Dev Opportunity Radar has been hearing from people who discovered something they otherwise might have missed.

You don't need to have been accepted or have a big success story to share. Sometimes simply discovering the right opportunity at the right time is already a win.

I hope this section gradually becomes a place where we can celebrate those stories together, one update at a time.

👋 Until Next Friday

Before I go, I just want to say thank you.

The goal has always been simple:

Help people discover opportunities they otherwise might have missed.

Thanks to all of you, it feels like we're doing exactly that.

Thank you for reading, for sharing, and for helping make this our radar, not just mine.

I'll be back next Friday with more opportunities, resources, and Community Finds.

Until then, take care, and I hope you discover something amazing this week 💙

🌐 Dev Opportunity Radar Website

👉 Website

I also wrote a short post about why I built the website and the journey behind it.

👉 Website Launch Post