Mithun Chavan A — Full-Stack AI Engineer

I am a full-stack AI engineer based in Bangalore, India. I build AI agents, developer tools and full-stack web applications that ship to real users. My current focus is on autonomous coding agents and browser-based AI code editors. I work with React, Next.js, TypeScript, Python and large language models every day.

About

I like building things that actually ship. Mostly AI agents, developer tools and web apps I can put in front of real people. I have shipped over 30 public repositories, built more than 15 full-stack and AI applications, and have 5 or more products live on the web across 8 or more languages and frameworks. I am passionate about creating software that solves real problems, from autonomous coding assistants that help developers write better code faster, to adaptive learning platforms that personalise education for every student.

Selected Work

Crayon — AI Coding Agent (2026)

Crayon is an AI coding agent that lives in your terminal and VS Code. It plans the work, runs the commands, and fixes its own broken builds. It indexes your codebase locally and lets you swap between Claude, GPT and local Ollama models. Built with TypeScript, Ink and LLM agent orchestration patterns. Crayon uses tree-sitter for code analysis and supports multi-file editing with automatic error recovery.

DevForge — AI Code Editor (2026)

DevForge is an AI code editor that runs entirely in your browser. It spins up a real Node environment using WebContainers, generates whole applications with live preview, and needs no backend at all. Built with Next.js, React 19, TypeScript and WebContainers. DevForge lets you go from a prompt to a running application in seconds, all without leaving your browser tab.

AdaptIQ — Adaptive Learning Platform (2026)

AdaptIQ is a maths practice app that adapts to each student using Item Response Theory. It reads how the student is doing and adjusts the difficulty in real time, drops in hints when they get stuck, and tracks progress across NCERT topics. Built with TypeScript, IRT algorithms and AI tutoring. The platform provides personalised learning paths and detailed analytics for both students and teachers.

InterPrep — AI Mock Interviews (2025)

InterPrep lets you practice interviews out loud with an AI. It runs technical and behavioral rounds by voice, scores you right after, and keeps every session so you can look back on how you did. Built with Next.js, Voice AI and Firebase. InterPrep covers data structures, algorithms, system design and behavioral interview questions.

Capabilities

My core skills include Agentic AI and autonomous agent design, LLM integration and prompt engineering, React and Next.js front-end development, TypeScript and JavaScript, FastAPI and Python back-end development, WebContainers and browser-based runtimes, GSAP and Framer Motion animations, Supabase and Firebase databases, and Computer Vision with OpenCV. I also work with OpenRouter, Ollama, Tree-sitter, SQLite, Tailwind CSS, WebSockets, Vite, Node.js and REST APIs.

Experience

Member of Technical Staff (MTS-1) at Aquera Labs, Bangalore, India (May 2026 – Present) — Working on AI-powered developer tools and full-stack product engineering.
SDE Trainee at Aquera Labs, Bangalore, India (Nov 2025 – May 2026) — Built and shipped production features across the web platform using React, TypeScript and Node.js.

Connect

Find me on social media and coding platforms: