Rupesh Bharambe
I build |
AI/ML Engineer · 6x Hackathon Winner · NFPC National Champion
Featured Projects
// hand-picked highlights
Mule Account Detection Pipeline
Problem: Money mule accounts launder illicit funds through legitimate-looking bank accounts, making fraud detection extremely difficult at scale. Solution: Two-pipeline stacking ensemble (LightGBM + CatBoost + XGBoost) with 615 engineered features, DuckDB streaming over 400M transactions on 16GB RAM, and rank blending for red herring resistance. Impact: National Champion — 0.999 AUC-ROC, 0.91 F1, beat 2000+ participants across 4 phases. Won ₹5,00,000 first prize.
Real-Time Electricity Demand Forecasting
Problem: Delhi's power grid lacked accurate short-term demand forecasting. Solution: Built SARIMAX + Transformer hybrid model on 12+ years of historical data enriched with real-time weather. Impact: 98%+ accuracy at 5-minute granularity, deployed on GCP with auto-retraining pipeline.
Face Liveness Detection System
Problem: Facial recognition systems are vulnerable to spoofing attacks (photos, videos, masks). Solution: Custom neural network with 477 facial landmarks + mathematical formulae for dynamic response-based liveness. Impact: 99% accuracy at <200ms latency, optimized for low-resource devices — applicable to KYC, banking, proctoring.
About Me
// quick intro
I'm a final-year B.Tech student in AI & Data Science, passionate about building intelligent systems that solve real-world problems. From winning Smart India Hackathon 2024 to building SatSage for ISRO's satellite data, I thrive at the intersection of research and production engineering.
Read more →