// developer · qa · ai-augmented
Developer and QA engineer. I ship Android apps, Python tools, JetBrains plugins and web experiments — and I lean on AI every day (Copilot, Claude, Codex, local LLMs, MCP agents) to move faster, test deeper, and stay sharp as the toolchain evolves.
A few projects from my GitHub.
AI-powered browser automation through MCP. Connect any MCP-compatible agent
(Copilot, JetBrains AI, Claude Desktop, Cursor…) to a real browser and let it
navigate, click, type, screenshot and inspect — guarded by safety tiers and
domain allowlists. Or skip the AI entirely: chat to a built-in
in-browser panel (⌘⇧K), or fire your own Playwright / Node
scripts through the extension.
A responsive multi-page restaurant site built with Django and deployed on Render.
Case study →Speech recognition and AI experiments in Python.
View on GitHub ↗A location-aware alerting application built for the IBM nalaiya thiran program.
View on GitHub ↗A simple YouTube video downloader utility written in Python.
View on GitHub ↗A demo Android app integrating Razorpay payments.
View on GitHub ↗Productivity plugins I've built for IntelliJ IDEA and the broader JetBrains family. Source lives in IntelliJ-Plugins.
Chromium/Firefox extension + Node MCP server that gives AI agents a real browser to drive — with layered safety tiers, local-only transport and token-efficient batch tools.
Manage Git's assume-unchanged flag from inside any JetBrains IDE —
with a dedicated tool window, multi-repo support and bulk actions.
Blocks git push when compilation errors exist — from both the IDE
push dialog and terminal git, via a managed pre-push hook.
Computers and gadgets have always piqued my curiosity. I started tinkering as a college student and I've been shipping small things ever since — mostly in Python, Java/Kotlin and Android, with a recent obsession for JetBrains plugin APIs.
I work across development and QA — building features one week, stress-testing and automating regression flows the next. That dual lens keeps my code honest and my test plans realistic.
I treat AI as a core part of my toolchain, not a gimmick. I've been an early beta tester of GitHub Copilot since 2022, back when it was invite-only — and I've watched (and helped shape, in my small way) how it's evolved ever since. Day-to-day I pair with GitHub Copilot, Claude and Codex for code, reviews and refactors; drive browser and IDE workflows through MCP agents (including my own AutoDOM); use LLMs to generate edge-case test data and explore unfamiliar codebases; and keep a healthy skeptic's eye on everything they produce. The goal is to stay relevant as the stack evolves — not to outsource the thinking.
“There are 10 kinds of people in the world: those who understand binary numerals, and those who don't.” — Ian Stewart
Have an idea, a project, or just want to say hi? My inbox is open.
aswin2kumarforme@gmail.com