DEV Community: Dhyey Patel

👻 PHANTOM//TRACE: Every Error Has a Ghost | AI-Powered Debugging Assistant

Dhyey Patel — Sat, 31 Oct 2026 18:22:39 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine

What I Built

Every developer has faced an error that seems impossible to understand.

That's why I built PHANTOM//TRACE, a spooky-themed AI programming error debugger.

Users can select a programming language, paste an error message, and ask the AI to investigate it.

The goal is to help developers understand:

What went wrong?

What is the likely root cause?

How can the error be fixed?

How can the same issue be prevented?

Inspired by the Ghost in the Machine theme, I imagined programming errors as mysterious ghosts hiding inside our code.

PHANTOM//TRACE helps developers uncover them. 👻

Demo

🚀 Live Demo: https://miyamoto-musashi11.github.io/PHANTOM-TRACE/

miyamoto-musashi11.github.io

Code

💻 GitHub Repository: https://github.com/Miyamoto-Musashi11/PHANTOM-TRACE

Miyamoto-Musashi11 / PHANTOM-TRACE

A spooky AI-powered programming error debugger using Hugging Face open-weight models.

https://miyamoto-musashi11.github.io/PHANTOM-TRACE/

View on GitHub

How I Built It

I built PHANTOM//TRACE using:

HTML — Application structure.

CSS — Spooky-themed responsive interface.

JavaScript — Frontend interactions and API requests.

Hugging Face Inference Providers — Access to AI models.

Qwen2.5-Coder-32B-Instruct — The coding model configured for error analysis.

The application sends the selected programming language and pasted error message to the Hugging Face chat completions API and displays the response.

Users provide their own Hugging Face access token for inference.

Why Does Open Innovation Matter?

Open innovation gives developers the freedom to experiment with AI models beyond closed ecosystems.

By building around an open-weight coding model, this project explores how accessible AI can help developers understand programming errors.

It also makes the application easier to customize, extend, and adapt to other compatible models.

I believe open innovation makes AI development more accessible to students, beginners, and independent builders.

My Agent Session

No DevRelay agent session recorded for this project.

The Spirit of It

Every developer knows that moment.

You write your code, run the program, and suddenly the terminal throws an error you have never seen before.

You investigate. You search. You try again.

Sometimes, debugging feels like fighting something invisible.

What if every error had a ghost?

PHANTOM//TRACE gives that ghost a voice.

Debug the unknown. Exorcise the error. 👻

Prize Categories

#devchallenge #hf26challenge
#GitHubCopilot #Gemma

LazyLoop AI: Let AI Handle the Planning | Hacktoberfest Week 3

Dhyey Patel — Sun, 25 Oct 2026 16:18:57 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady

What I Built

LazyLoop AI — Let AI Handle the Planning

LazyLoop AI is an AI-powered task planning assistant that helps users get started on tasks without spending too much time figuring out what to do.

Users enter a task or goal, and the application sends it to an AI model to generate a helpful response. The idea is simple: reduce the effort required to get started and make everyday work a little easier.

I built this project to explore how open-source AI can be integrated into a practical web application.

Demo

GitHub Repository: https://github.com/Miyamoto-Musashi11/LazyLoop-AI

Video Demo: [Add your demo video link here]

The demo shows the user entering a task, clicking Generate with AI, and receiving an AI-generated response.

Code

Source Code: https://github.com/Miyamoto-Musashi11/LazyLoop-AI

The repository contains the frontend, Node.js backend, and supporting services used by LazyLoop AI.

How I Built It

I built LazyLoop AI using:

HTML, CSS, and JavaScript for the frontend.
Node.js for the backend.
Ollama for local AI inference using an open-weight model.
Git and GitHub for version control and sharing the source code.

The frontend sends the user's task to the backend through an API endpoint. The backend processes the request and communicates with the AI model through Ollama before returning the generated response.

Building this project helped me understand how a frontend, backend, and locally running AI model can work together in a real application.

Why Does Open Innovation Matter?

Open innovation makes AI development more accessible to students and independent developers like me.

Using Ollama with an open-weight model gave me an opportunity to experiment with AI inference locally instead of relying entirely on a paid, closed AI API.

It also helped me learn by building, testing, and improving a working application while keeping the underlying technology more flexible.

For me, open innovation means having the freedom to experiment, learn, and create useful tools without needing an expensive setup.

My Agent Session

I developed and tested LazyLoop AI through hands-on coding and debugging, including getting the API request working and verifying that the application returns generated responses.

Optional: Add a DevRelay agent session link here if you have one.

The Lazy Payoff

LazyLoop AI aims to reduce the friction of getting started on a task by turning a user's request into an AI-generated response.

Rather than starting with a blank page, users can describe what they need and use the generated response as a starting point.

The current version successfully returns AI-generated responses. Improving response speed and making generated code more reliable are the next steps on my roadmap.

Prize Categories

Add only the partner prize categories that apply to your project and that you intend to enter, based on the official challenge rules.

Built with curiosity, open-source tools, and a little laziness. 🚀

hf26challenge #devchallenge #OpenSourceAI #Ollama #NodeJS

GigPitch AI: Turn Your Skills Into Freelance Opportunities With Open-Weight AI

Dhyey Patel — Sun, 18 Oct 2026 18:23:19 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

GigPitch AI is a prototype designed to help beginner freelancers turn their skills into potential side hustles.

It aims to help users discover freelance service ideas, suggest pricing, draft client proposals, and prepare outreach messages. The target users are students, beginners, and aspiring freelancers who want to find their first clients but struggle with presenting their skills professionally.

The idea is simple: make starting a small freelance business easier with AI assistance.

Demo

The project currently runs locally. A public demo is not available yet.

GitHub repository: https://github.com/Miyamoto-Musashi11/GigPitch-AI

Code

Miyamoto-Musashi11 / GigPitch-AI

An AI-powered freelance side-hustle assistant using open-weight AI.

GigPitch-AI

An AI-powered freelance side-hustle assistant using open-weight AI.

View on GitHub

How I Built It

I built GigPitch AI using HTML and CSS for the frontend, JavaScript for user interactions, and Node.js with Express for the backend.

I chose Ollama to run the open-weight Qwen 2.5 3B model locally. The application is designed to send user requirements to the model and generate freelance service ideas, pricing suggestions, proposals, and outreach messages.

The project is an early prototype, and the complete AI workflow still needs testing.

Why Does Open Innovation Matter?

Open innovation makes AI development more accessible to students and independent developers who may not be able to afford paid API services.

Using open-weight models with local inference gives developers more control over their tools, creates opportunities to experiment, and can reduce dependence on proprietary AI platforms. It also makes learning about AI integration more accessible to beginners building practical projects.

My Agent Session

I used AI-assisted development to build the GigPitch AI prototype, including its frontend, Node.js backend, and planned local integration with Ollama and Qwen 2.5 3B. A saved DevRelay session is not available for this submission.

Prize Categories

Side Hustle

#hf26challenge #devchallenge

This is an individual project. I built GigPitch AI independently.

My first DEV Post

Dhyey Patel — Wed, 14 Oct 2026 17:04:33 +0000

🚀 Starting My Developer Journey: Learning, Building, and Growing

Hey everyone! 👋

I'm a B.Tech student passionate about software development, Artificial Intelligence, and Machine Learning.

I'm currently exploring web development, Data Structures and Algorithms, and new technologies while working on projects to improve my practical skills.

I believe the best way to learn programming is by building projects, making mistakes, and learning from the developer community.

🎯 My current goals:

Improve my coding and problem-solving skills.
Build useful real-world projects.
Explore AI/ML and modern web technologies.
Start contributing to open-source projects.
Connect with developers and learn from their experiences.

I'm excited to be part of the DEV Community and looking forward to sharing my progress, projects, challenges, and lessons along the way.

I'm still learning, but I'm ready to build and grow! 💻✨

I'd love to hear your advice: What is one thing you wish you knew when you started your developer journey?

beginners #webdev #programming #learning