Currently building an autonomous stat-arb bot live on NSE markets and doing clinical ML research on correlated multi-disease detection.
- Autonomous systems that run 24/7 with real money on the line
- Statistical rigor first — if the math doesn't hold, the code doesn't ship
- Final year @ KIIT · IEEE Student Researcher
| Project | What it does | Result |
|---|---|---|
| Iron-Sentry | Autonomous NSE pairs trading bot — OLS hedge ratio, z-score mean reversion, ghost-order guard, Telegram alerts | ≥1 trade/day guaranteed, live paper trading |
| ComorbidNet | Detects T2D + HTN + MetS + CKD simultaneously via Classifier Chains — captures the metabolic cascade independent models miss | +0.03 AUC gain, Hamming loss 0.09 → 0.07 |
| CreditSense | Real-time credit risk API — LR + RF + GBM ensemble, FastAPI backend, one-click Render deploy | Trained on 23K+ real loan records |
| V2V Perception | V2V collaborative hazard detection — DeepLabV3 segmentation → XOR delta zones → 48K CNN decision model | HACK4IMPACT Track 2 submission |
Statistical Arbitrage · Classifier Chains · SHAP · XGBoost · OpenCV
asyncio · DeepLabV3 · NLP · Transformers · Power BI · Grafana · Redis