The filled squares are my 357 commits, coloured by project and placed in the order I made them. The outlined ones are still empty. Hover a filled one.Tap a filled one.
An AI code reviewer that finds bugs, security holes and complexity problems, then hands back the fixed code.
2 · analysis results
3 · code explainerI grew up drawing. Now I build full-stack and AI products, some in a day, one in twelve, and I care how they look as much as how they work.
I'm a second-year CSE student at MSU Baroda. I still make things the way I learned to draw: a rough sketch first, then refining until it looks right. These days the canvas is full-stack and AI products.
The heavy ones I ship with. The light ones, I'm still learning.
Hover a skill, or a projectTap a skill, or a project
My journey: first commit in December 2025; CodeAutopsy, my first project; Clarity for TetraTHON's entry round; top 44 of 160 at TetraTHON; Prism in two days; GlobeTrotter at the Odoo hackathon; Swasthya-AI built solo; a Decode SIH certificate; and now Smart India Hackathon 2026, where we cleared the college round and our prototype is under SIH review.
$ git tag -a sih-2026 -m "Smart India Hackathon 2026: cleared college round"
Team Caffeine Coders, six of us. Problem statement 26041 from the Government of Jharkhand: an AR-based vocational training simulator for industrial safety in mining and manufacturing. Our prototype, Suraksha Saathi, is a mobile app with a FastAPI back end and a React dashboard.
$ git tag -a decode-sih -m "Decode SIH: certificate"
Earned during Smart India Hackathon 2026.
$ git tag -a tetrathon-2026 -m "TetraTHON 2026: top 44 of 160"
We entered with Clarity, a credit-likelihood and investment planner, and made the top 44. For that round we built Prism, an AI diligence tool, in two days. We weren’t selected for the top-9 pitch.
A credit-likelihood score and micro-investment planner for people banks usually can’t score.
Millions of people in India’s Tier-2 and Tier-3 cities have little or no credit history, so bureaus can’t score them and lenders turn them away. They are also rarely shown how small, regular investments could grow.
A React front end on a FastAPI back end. The score is deliberately interpretable: a weighted average of normalised answers mapped to 300–900, with no protected attributes, synthetic data only, and designed with India’s DPDP Act 2023 in mind.
I built and shipped the entire front end: the chapter-based interface, questionnaire, score reveal, explainability views, sliders and charts. I also fixed the input bugs found in testing, wrote the README and set up analytics. Vihan built the back end.
Our entry-round submission for TetraTHON 2026. It put us in the top 44 of 160 teams.