SPIT Mumbai · 9.69 CGPA
Building intelligent systems at the intersection of ML and full-stack engineering
B.Tech student at SPIT Mumbai building production-grade ML systems and full-stack applications. Currently focused on machine learning, real-time systems, and making AI models explainable and useful.
- 🎓 Electronics & Telecom Engineering · 9.69 CGPA
- 🏆 Amazon ML Summer School 2026 — national shortlist
- 🌐 GSSoC 2026 Legend — 2196 pts, 24 merged PRs across 5 repos
Stockfish tells you the best move. ChessTrace tells you what you'll actually play — and when you'll crack.
chess-trace.vercel.app · GitHub
A chess intelligence platform that models human play under time pressure — going beyond Stockfish and Chess.com by decoding each player's unique cognitive fingerprint.
What makes it different:
- 8 XGBoost models trained on 512K real move plies with clock data
- Real-time WebSocket live game spectating with sub-2s ML inference
- SHAP plain-English explanations surfacing psychological + time-pressure factors
- Personal cognitive fingerprinting: tilt detection, accuracy vs. clock degradation curves, K-Means archetype clustering, GM style matching via cosine similarity
- Inspired by Maia Chess (Cornell + Microsoft Research, KDD 2020)
Python FastAPI XGBoost SHAP React WebSockets Stockfish PostgreSQL
Languages
ML / Data
Backend / Full-Stack
