DEV Community: Muskan Fatima

8 Hours, One Image Bug, and a Lesson in AI Development

Muskan Fatima — Mon, 02 Mar 2026 10:36:44 +0000

This is a submission for the Built with Google Gemini: Writing Challenge

What I Built with Google Gemini

For the New Year Portfolio Challenge, I built my personal AI-powered developer portfolio using Google Gemini inside AI Studio and deployed it directly to Cloud Run.

The goal was to:

Build a modern responsive portfolio
Showcase my AI, Web 3.0, and Metaverse projects
Use Gemini as a collaborative coding partner
Deploy without traditional local setup

Instead of writing everything manually, I used Gemini to:

Generate layout structure
Refine UI components
Improve responsiveness
Debug code issues
Optimize deployment

This wasn’t just code generation. It was AI-assisted development.

🔗 Live Demo:

What I Learned

AI Studio ≠ Local Development

My biggest challenge was image handling.

At first:

Gemini added random stock images.
I wanted to use real project screenshots.
I created a public folder manually.
I placed images there.
I updated image paths in code.

It didn’t work.

I spent nearly 8 hours debugging this.

After multiple failed attempts, I realized something critical:

AI Studio does not handle local assets the way traditional local environments do.

The solution:
Upload images online → Use public URLs.

The moment I switched to hosted image URLs, everything worked.

Lesson:
Understanding the environment matters as much as understanding the code.

Prompt Precision Changes Everything

When I asked:
“Fix the UI.”

Gemini responded:
“Fix completed.”

But the layout was still not fully responsive.

When I changed my prompt to:
“Make the UI fully responsive across mobile, tablet, and desktop using proper flex/grid adjustments.”

It worked.

AI is powerful — but clarity controls output.

Debugging AI-Generated Code Is Still Debugging

Even when AI says something is fixed, you must verify.

AI speeds up development.
It does not remove responsibility.

That mindset shift was important.

Google Gemini Feedback

What Worked Well

Fast UI scaffolding
Clean component structure
Rapid iteration
Direct Cloud Run deployment (huge advantage)
Smooth overall developer experience

The direct deployment feature was a game changer. No manual cloud setup saved a lot of time.

Where I Faced Friction

No clear warning about local image limitations
AI sometimes confirmed fixes that weren’t fully implemented
Required multiple prompt refinements for accurate responsiveness

If AI Studio could:

Detect unsupported local asset behavior
Provide environment-aware debugging hints
Validate UI responsiveness automatically

It would dramatically reduce friction for developers.

Looking Forward

This experience changed how I approach AI development.

Next, I want to:

Build more complex AI-integrated apps
Experiment with agent-based workflows
Explore production-grade AI systems
Contribute to AI-first developer tooling

The biggest transformation wasn’t the portfolio.

It was learning how to collaborate with AI intentionally.

AI is not magic.
It’s leverage — when guided properly.

From a 30-Minute AI Build to a Real Portfolio That Tells My Story

Muskan Fatima — Sun, 04 Jan 2026 13:54:11 +0000

This is a submission for the New Year, New You Portfolio Challenge Presented by Google AI

Before — Protofilo-2 (Static Portfolio)

After — AI-Enhanced Portfolio (Cloud Run + Real Images + AI Twin)

Before vs After — my portfolio’s evolution from static and generic to authentic and meaningful.

I built a portfolio in 30 minutes using Google AI Studio — fast, flashy, and deployable.
After submitting it, I realized something crucial:

Real visuals and real effort matter more than speed.

What happened next changed my mindset and the quality of my work.

The Problem I Faced

At first, I relied heavily on AI Studio — including AI-generated images for my portfolio.

Everything worked perfectly locally, but after deployment:

Images didn’t load
Mobile layout broke
Public and assets folders didn’t behave as expected

I tried fixing paths again and again, but nothing worked.

Then I learned an important technical detail:

AI Studio does not behave like a traditional local development environment.
Local image paths often break after deployment.

This realization completely changed my approach.

The Debugging Breakthrough

Instead of forcing local assets, I switched to cloud-hosted images.
Once I did that:

Images loaded correctly
Mobile responsiveness improved
Deployment became stable

I redeployed — this time with clarity and confidence.

Visual Transformation — Before vs After

Before — Protofilo-2

Static presentation
Generic layout
Basic project listing
No strong personal identity

After — AI-Enhanced Portfolio

Real images everywhere
Mobile-responsive layout
AI-powered components (AI Twin)
Clear storytelling and intention

This comparison shows my shift from a generic static portfolio to a meaningful, functional, and authentic one.

What I Learned

Lessons from the Challenge

AI isn’t a shortcut — it’s a tool

It speeds up development, but you still need to debug and think like an engineer.

Authenticity matters

Fake images create distance. Real visuals build trust.

Deployment teaches real engineering

Cloud environments behave differently than local setups.

Work smart, not just fast

The right tools + correct approach save time and energy.

How AI Studio Helped Me

Live Portfolio

👉 Live URL:
https://muskan-fatima-portfolio-1075074220803.us-west1.run.app

Final Reflection

This submission isn’t just code — it represents my growth.

I started with speed.
I ended with purpose.

This portfolio now tells my story — honestly, technically, and confidently.

I’m excited for what comes next, because now I build with intention, not just speed

Turn Any YouTube Lecture into Simple Urdu – EduBridge Urdu

Muskan Fatima — Wed, 10 Dec 2025 07:13:54 +0000

Title:

EduBridge Urdu – Turning Any YouTube Lecture into Simple Urdu for Students

My Pitch Video

stream.mux.com

Demo

Live Demo / Walkthrough: https://drive.google.com/file/d/1iWnptkPNzH2tWM-1O2HKe_jrn5csI2GS/view?usp=sharing
Testing: Paste any YouTube lecture link

The Story Behind EduBridge Urdu

Hi, I’m Muskan!
EduBridge Urdu was born from a simple but urgent problem:

Millions of Pakistani students want to learn from the world’s best YouTube lectures… but English becomes a barrier.

Students struggle not because the topics are hard — but because the language feels impossible.

So I built EduBridge Urdu to remove that barrier instantly.

Paste any YouTube lecture link →
You get the full lecture translated into simple, WhatsApp-style Urdu, plus a clear Urdu audio voice-over.

Students simply mute the original video and play the Urdu audio.
For the first time, they understand the entire lecture effortlessly.

What started as a small idea became a real mission:
Make global education understandable for every student, no matter their English level.

It wasn’t easy.
We tried multiple APIs, rebuilt the system twice, and hit many accuracy challenges.
But eventually, with a Paid Gemini API AI and more research, we achieved a smooth, student-friendly Urdu learning experience.

EduBridge Urdu isn’t just a translator — it’s a bridge to global knowledge.

Technical Highlights

Full AI agent pipeline built from scratch
Complete lecture translation, not summaries
Urdu text + natural Urdu voice generation
Clean, conversational language — easy for all ages
Optional Hindi-English summary
Scalable for long YouTube lectures
Frontend built by teammates; AI system fully developed and integrated by me
Built specifically for real student challenges in Pakistan

⭐ Use of Mux (Best Use of Mux Category)

Mux was used to host and stream my pitch video.
Its automatic pipeline handled:

Adaptive bitrate streaming
Smart encoding
Playback optimization
Reliable delivery across devices

I didn’t touch complex video infrastructure — Mux handled it beautifully.

Why This Project Matters

EduBridge Urdu opens access to global education:

✔ Students can learn from Harvard, MIT, YouTube educators
✔ Non-English speakers finally understand advanced topics
✔ Schools, coaching centers, and self-learners benefit
✔ It reduces stress and builds confidence

This can scale to millions of learners across Pakistan and beyond.

⭐ Thank You!

Excited to share this with the DEV community.
If you’d like to try it or give feedback, I’d love to hear from you!

How the AI Agents Intensive Changed My View of SDKs and Agent Development

Muskan Fatima — Thu, 04 Dec 2025 13:42:01 +0000

This is a submission for the Google AI Agents Writing Challenge: Learning Reflections

I wasn’t even planning to join any course, but one day while scrolling Instagram, a reel about the AI Agents Intensive appeared on my feed. I checked it quickly and enrolled on the spot. I was excited because it was officially from Google, and it introduced a new SDK that I had never explored before. I thought it would be a great chance to earn a Google certificate and learn something fresh.

Even though I had experience with the OpenAI SDK and a few others, I had never worked with MCP inside an SDK. At first, I assumed this SDK would work like the others. I was completely wrong. It may look similar, but it’s totally different — in a powerful way. It’s the only SDK I’ve seen that offers built-in MCP connectivity, free Google Search tools, and such clean session + context management. Everything felt organized and understandable.

I’m really thankful to the entire team, especially the person who created the notebooks. Every detail was explained so clearly that it genuinely changed my view about what an SDK can do. I didn’t expect this level of efficiency and simplicity.

As a student and beginner with this SDK, I tried to push myself. I even attempted to build a frontend using Chainlit and Streamlit around it. I failed almost 20 times. Nothing worked the way I wanted. One day before the final deadline, I suddenly thought: why not wrap everything with FastAPI like I did with other SDKs? I tried it — and it worked perfectly. I connected the SDK with FastAPI and a simple index.html frontend, and the result looked fantastic. That moment felt like a big win.

This intensive didn’t just teach me a new SDK — it changed how I think about agents, tools, and automation. My next goal is to build a complete project using this SDK, automation, and Next.js, and create something that surprises everyone. I’m excited for what comes next.

Getting to Know Gemini CLI: Your Terminal’s New AI Ally

Muskan Fatima — Wed, 05 Nov 2025 17:09:39 +0000

Whether you’re building a frontend interface, automating tasks, or just exploring how AI can ease your workflow — Gemini CLI is worth a look.

Introduction

For many frontend developers, most of our day is spent in the browser, code editor, or terminal.
What’s exciting about Gemini CLI is that it brings a powerful AI assistant directly into your command line — where we already live most of the time.

💡 Google describes Gemini CLI as “an open-source AI agent that brings the power of Gemini directly into your terminal.”

It’s not just another chatbot — it understands your code, helps debug and generate, and even connects to custom tools.

💡 Why Gemini CLI Was Built

Here are a few reasons why this tool was created:

🧠 Developers spend tons of time in terminals and IDEs. Bringing AI there means less context switching.
🧩 Generative AI should be part of workflows — not just “chat in a window,” but “tools we call from scripts.”
⚙️ Flexibility to write code, fix bugs, generate content, or manage tasks — all from the CLI.

Brief History & Key Milestones

June 2025: Official launch by Google.
License: Open-source under Apache-2.0.
Early Features:
- Large context window (1M tokens)
- Free developer tier (60 requests/min, 1,000/day)
Later Additions:
- Custom slash commands via .toml files
- IDE companion and Model Context Protocol (MCP)

⚡ What Makes Gemini CLI Powerful

Here’s why developers are excited about it:

*Terminal-first design:*Works on Windows, macOS, and Linux — no heavy UI needed.

Integrations:*Supports shell commands, file ops, and web fetching.
MCP support:* Extend it with your own servers/tools.
Project-aware context: Reference files, directories, or even images and PDFs.
⚡ Custom commands: Create /review, /plan, or any slash command using .toml.

Getting Started (Especially for Frontend Devs)

Here’s a quick setup guide:

1️⃣ Install

Make sure you have Node.js v20+ installed.

npm install -g @google/gemini-cli

2️⃣ Authenticate

You can log in using your Google account or an API key.

Option 1: Google Login

Requires a Gemini Code Assist license (free tier available).

Option 2: API Key

setx GEMINI_API_KEY "your_api_key_here"

or on macOS/Linux:

export GEMINI_API_KEY="your_api_key_here"

3️⃣ Run Gemini

gemini

You’ll enter an interactive session — ready to chat, code, or automate.

Try Something Practical

Ask Gemini to explain a component from your codebase:

> Explain how this React component works. @src/components/Button.jsx

Or create your own custom slash commands:

# .gemini/commands/plan.toml
[command]
description = "Generate a project plan"
prompt = "Create a detailed plan for this project"

Then simply type:

/plan

Future Directions

Expect Gemini CLI to evolve with:

Deeper IDE integrations
More community-built MCP servers
Integration in CI/CD workflows (AI code review, automation)
Stronger connections with agent frameworks and metaverse tools

Conclusion

In short, Gemini CLI bridges the gap between AI models and your everyday dev workflow.

If you’re a frontend developer exploring new tech like AI and automation — this tool will fit perfectly into your learning and building journey.

💭 Try it. Experiment. Build. And most importantly — have fun with it!

3 Unfair Advantages Pakistani Developers Have in GSSOC (And How to Weaponize Them)

Muskan Fatima — Fri, 18 Jul 2025 08:27:40 +0000

Why your Karachi internet struggles and Lahore sleep schedule might be secret weapons for open-source domination."

🔥 Advantage 1: The Resilience Multiplier

Global Myth:

"Stable internet and 24/7 coding time are necessary"

Pakistani Reality:

Our daily struggles - load shedding, ISP issues, university pressure - force you to master:

Hyper-focused coding sprints (2 hours of electricity? Crush 1 PR)
Offline development (Git workflows without constant internet)
Resourcefulness (Solving problems Stack Overflow can't answer)

Actionable Tactic:

"The Power Cut Protocol"

Clone repos during stable internet

Work locally using git add -p for surgical commits

Batch PR submissions during peak connectivity

(Example: Ahmed from Peshawar merged 8 PRs using 1GB mobile data)

🌙 Advantage 2: Time Zone Jiu-Jitsu

Global Problem:

Most contributors work 9AM-5PM in their timezones → PR rush hours = intense competition.

Your Secret:

6PM PKT = Golden Hour

Indian maintainers are awake (5:30PM IST)
European maintainers starting day (3PM GMT)
US West Coast maintainers just waking up (6AM PST)

Actionable Tactic:

"The Twilight Contribution Strategy"

5-7 PM PKT: Submit PRs → Avoid daytime traffic

10 PM PKT: Review PR comments from Europe/US (My GSSOC dashboard showing 73% PR merge rate at 6PM PKT)

💡 Advantage 3: The Underdog Network Effect

Global Blindspot:

Most contributors compete individually.

Your Weapon:

Pakistan's emerging tech ecosystem means:

Less competition for local-first issues (urdu pakistan localization tags)
Stronger community bonds (WhatsApp groups > Discord servers)

Actionable Tactic:

Create a "Shadow Squad":

Find 3 Pakistani GSSOC participants

Assign roles: Researcher, Coder, Documenter

Raid good-first-issues as a team (Group from LUMS merged 15 PRs in Week 1 using this)

🚀 Live Deep Dive: Saturday 6PM PKT

I'm hosting an exclusive session:

"GSSOC Warfare: Pakistani Edition"

→ Leaked: Maintainer preference data for PKT timezones

→ Live Demo: Offline Git workflow for load-shedding scenarios

→ Blueprint: Building your "Shadow Squad"

Claim Your Spot:

https://us05web.zoom.us/j/6625456294?pwd=bb7ccyZhOxoRsYgrXigapF6z41I2Bz.1&omn=87282206455

🇵🇰 Why This Matters Now

We're at an inflection point:

2023: 12 Pakistani GSSOC winners
2024: 27 and counting
2025: Your turn.

The global OSS scene needs our resilience. Let's show them what Karachi nights and Islamabad grit can do.

"You don't need better conditions. You need better strategies."

Comment below with your city! Let's connect 👇

How GSSoC Helped Me Build My Portfolio as a Frontend Developer

Muskan Fatima — Sat, 12 Jul 2025 09:11:15 +0000

✨ How GSSoC Helped Me Build My Portfolio as a Frontend Developer

Open source was always something I admired from afar — a space where developers from all over the world contribute, learn, and grow. But it wasn’t until I became a contributor in GirlScript Summer of Code (GSSoC) that I truly realized how powerful this journey could be.

Real Work, Real Impact — Without a Traditional Internship

Before GSSoC, I had no traditional internship on my resume. Still, I wanted to gain real-world development experience. Through GSSoC, I contributed to actual, live open-source projects that were being used and built by others. These weren’t just tutorials — they were collaborative efforts where my code made an impact.

This gave me credible work experience that now shines on my portfolio, and recruiters have taken notice!

Stronger Portfolio, Better Opportunities

Every project I contributed to during GSSoC added something unique to my GitHub:

Pull Requests on real repositories ✅
Issue tracking & problem solving 🧠
Clean, readable code that others rely on 🧹

As a frontend developer, GSSoC helped me put theory into practice — working with React, Tailwind CSS, APIs, and collaborative workflows.

GitHub Growth & My First Badge!

One of the most exciting moments? I earned my first GitHub badge — Pull Shark 🦈, for consistently submitting pull requests during the program.

GSSoC also brought more visibility to my GitHub profile:

More followers 👥
More stars ⭐
More engagement from the community 🔁

💬 Final Thoughts

Thanks to GSSoC, I didn’t just build a stronger portfolio — I built confidence, connections, and real experience. If you’re a student or beginner developer looking to take that first big leap into open source, GSSoC is the best place to start.

You don’t need a fancy internship.
You need opportunity — and GSSoC gave me that.

Feel free to check out my GitHub or twitter to connect with me and learn more about GSSoC!

🔍 Do Agents Call Tools Every Time? Here's the Truth and How to Control It

Muskan Fatima — Mon, 16 Jun 2025 07:45:40 +0000

🧠 Do Agents Call Tools Every Time?

❌ No, not always.

If your agent is built correctly using frameworks like agents, LangChain, or OpenAI’s Assistants API, the model decides whether to call a tool — based entirely on the user input and intent.

Just defining a tool doesn’t mean it will be called every time.

✅ Example:

tools = [get_weather, get_muskan_data]

🟢 If the user says: "What’s the weather in Lahore?"
→ Only get_weather() will run.

🟢 If they ask: "Tell me about Muskan Fatima"
→ Only get_muskan_data() will be triggered.

This is what we call selective tool invocation — and it’s handled automatically by the AI model.

🔧 Why Do Tools Sometimes Feel Like They’re Always Being Called?

That could happen if:

❌ You’re manually calling the function every time (e.g., tool() in code).
❌ You're not using a model that supports tool-calling properly.
❌ The framework isn’t giving the model the option to choose — tools are hardcoded in logic.

✅ How to Control Tool Calling

1. Let the Model Decide (Default Behavior)

When using agents or OpenAI SDKs:

Agent(
  name="MyAgent",
  tools=[get_weather, get_news],
  instructions="You're an assistant helping with daily updates."
)

✔️ The model will automatically decide which tool (if any) to use.

2. Use Tool Control Flags (Advanced)

In some frameworks (e.g., OpenAI Assistant API), you can explicitly control tool behavior:

Agent(
  tools=[get_weather],
  tool_choice="auto"  # Model decides based on intent
)

Other options include:

Option	What It Does
`"auto"`	✅ Model decides whether to call a tool
`"none"`	❌ Tools will not be called at all
`"required"`	⚠️ Forces tool use (not common)
`"tool_name"`	🔧 Force a specific tool to be called

🛠 Want Even More Manual Control?

You can add logic to prevent tool execution:

def tool_wrapper(func):
    def safe_call(*args, **kwargs):
        if should_call_this_tool():  # your own logic
            return func(*args, **kwargs)
        return "Tool not needed right now."
    return safe_call

This lets you override the AI's judgment if needed.

🔍 Summary Table

❓ Do tools get called every time?	❌ No — model decides based on prompt/input
🛠 Can I stop them from being called?	✅ Yes — using `"tool_choice"` or custom wrappers
📦 Should I define all tools upfront?	✅ Yes — but they will only be used when needed
🚨 Can I force-disable them globally?	✅ Yes — use `"tool_choice": "none"`

🎨 Suggested Image Prompt for Dev.to Post

"A digital illustration of an AI agent at a control panel with different tools (like a weather app, profile info, etc.) on the screen. Some tools are glowing while others are dimmed, representing selective tool calling. Clean, modern style with a tech/AI theme."

Use this prompt in any AI image generator like DALL·E or Gemini.

Final Thought

Tool calling is what makes AI agents truly useful. But understanding how to control it makes you a smarter engineer.

If you’ve ever asked yourself:

“Why is this tool running again?”

Then congrats — you’re already thinking like a pro agent developer. 🔥

📣 Bonus Question for Readers

✨ Have you ever accidentally triggered a tool in your agent that you didn’t need? How did you fix it?
Let’s chat in the comments!

“Prompt Design Is an Art — Here's the Brushstroke Guide”

Muskan Fatima — Sun, 15 Jun 2025 15:47:23 +0000

Prompt design patterns are structured ways of writing prompts to guide an AI's behavior effectively.
There are three main types:

Role-Based Prompt
Task-Oriented Prompt
Output-Guided Prompt

1️⃣ Role-Based Prompt

🧠 What it is:
In a role-based prompt, you assign the AI a specific role or identity, which influences how it responds.

✅ Example:

You are a content writer. Your task is to help users create blog posts, taglines, and product descriptions.

💡 Fun Fact:
This pattern helps the AI “think” like an expert in a specific domain, improving its tone and content relevance.

2️⃣ Task-Oriented Prompt

🧠 What it is:
You clearly define the task the AI must complete, without necessarily assigning it a role.

✅ Example:

Create a resume with a clean layout and relevant sections.

💡 Fun Fact:
Task-oriented prompts are the most common in automation and coding use-cases, where the "what" matters more than the "who."

3️⃣ Output-Guided Prompt

🧠 What it is:
You specify both the task and give clear guidance or format for how the output should look.

✅ Example:

Design a modern resume that looks like this:

Brown sidebar on the left

White background for the main content

Title at the top

Profile icon centered in the sidebar

Minimal and clean section layout

If prompt design is an art… what masterpiece are you secretly waiting to create? Share it in the comments!

Prompt Like a Pro: Zero-Shot, Few-Shot & Chain-of-Thought Explained Simply

Muskan Fatima — Sat, 14 Jun 2025 11:19:01 +0000

What is Prompting?

A prompt is the input we give to an AI agent — it's a combination of instructions and context that helps the AI understand what kind of output we expect. The better and clearer your prompt is, the better and more relevant the response you'll get from the AI.

When working with AI models (like chatbots or generative tools), prompting becomes a powerful skill. There are three main types of prompting:

🟢 1. Zero-Shot Prompting

In zero-shot prompting, we ask the AI a question directly without giving it any examples. We're expecting it to understand and generate a response based solely on its existing knowledge.

Example:
"Write a cover letter for a frontend developer job."

🟡 2. Few-Shot Prompting

Few-shot prompting means we give the AI a few examples of what we want — usually 1 to 3. These examples help guide the AI toward producing better results.

Example:
You show the AI a design or give it 1–2 samples and say:
"I want something like this, with these features."
You're giving it a reference to follow.

🔵 3. Chain-of-Thought Prompting

Chain-of-thought prompting is used when we want the AI to reason step by step. Here, we often give more examples (sometimes 5 to 9 or more), and we also ask the AI to think through the problem logically before answering.

Example:
“Let’s think step by step...”
or
“First we calculate X, then we find Y...”

This is especially helpful for complex reasoning, coding, or multi-step tasks.

👀 Before reading this, did you know there were different types of prompts? Which one was new to you?

My Story of Failure

Muskan Fatima — Fri, 06 Jun 2025 14:58:34 +0000

Today, I’m not here to share a new tech tutorial or a coding trick. Instead, I want to open up about something personal—a recent experience that didn’t go the way I hoped. It’s a story of failure, but also of growth, learning, and strength.

📉 The Rollercoaster of Job Hunting

For the past few months, I’ve been actively applying for jobs—consistently, persistently, and with hope in my heart. Last week, something unexpected happened: I received two job offers on the same day—one from Condlingo and another from Aptech. That was the happiest day for me! I felt secure, believing that if one didn’t work out, the other surely would.

📶 The Internet Glitch That Cost Me

My first interview was with Condlingo. I was well-prepared, excited, and ready to give my 100%. But life had other plans.

Right when it was my turn, my internet started acting up. To make matters worse, the electricity went out. I tried to stay calm, but eventually, the interviewer told me to disconnect. I messaged her afterward, requesting another chance, but I never got a reply. That silence hurt more than the disconnection.

🎤 My First Offline Interview – A Confidence Hit

I still had hope—Aptech was my second chance. I went for their in-person interview, which included both a verbal interview and a live presentation. The initial one-on-one session went well.

But when it came time to present in front of their team, something changed. My nerves took over. I felt underconfident. I forgot things I knew. I messed up. Walking out of there, I felt like I’d blown it.

🌱 But Here’s the Truth

I’ve realized that first times are rarely perfect. We all have those moments when things don’t go as planned, when life tests us hard. But we are not defined by one bad interview or one missed opportunity.

What matters is what we do next.

I’ve promised myself: I won’t beg for a second chance—I’ll become the version of myself that companies won’t want to miss next time. I’ll work on my weak areas, build my confidence, and keep learning.

❤️ To Anyone Who's Failing Right Now

Don’t give up. Every failure is a stepping stone, every setback a lesson. We’re all on a journey, and nobody succeeds without falling a few times first.

Fail. Feel. Learn. And then—rise stronger.

Do you also face a failure that hit you tell me in comments 🙃

🧠 Difference Between pydantic and mypy

Muskan Fatima — Sun, 18 May 2025 06:12:22 +0000

🧩 Summary for Developers

Use Pydantic when you need to validate and clean data at runtime, especially in APIs or when accepting user input.

Use mypy when you want to catch type-related bugs early during development through static type checking.

Both tools improve code quality but serve different purposes:
➡️ Pydantic keeps your data clean while running.
➡️ Mypy keeps your types clean before running.

Use them together for best results in modern Python projects.