Full-Stack AI Engineer building agentic systems, RAG pipelines, and real-time computer vision products that actually ship.
I grew up in Gurgaon, surrounded by glass towers that touched the sky — monuments to ambition and relentless engineering. I used to look up and wonder what invisible systems made them alive. TNow, I build the intelligence inside them.
Currently engineering AI-driven compliance automation at Complai — end-to-end RAG pipelines that process complex legal documents for FinTech. Previously, I led computer vision at Smart Eye , shipping YOLO inference on Raspberry Pi for edge deployment.
I've entered 4 hackathons and won all 4 — from a first-timer's nerves in Mumbai to a 14-day enterprise sprint in Hyderabad that ended with a crystal trophy and a job offer. I also authored a research paper on credit risk drift, presented at the STELMAR Conference.
Engineering AI-driven compliance models for regulatory risk assessment. Developing end-to-end RAG pipelines with advanced chunking and vectorization to process complex legal documents, drastically reducing manual audit time.
Architecting backend and AI integration for an enterprise-grade Active Learning Platform. Microservices for video transcription, automated reporting, and context-aware LLM workflows.
Led engineering of a smart blind stick using YOLO-Tiny on Raspberry Pi. Built async pipeline processing live video frames and TTS concurrently for zero-latency edge inference at 30+ FPS.
4-tier microservices platform orchestrating LLM-driven video summarization, mind-map generation, and automated assessments. Browser-native Edge Monitor uses MediaPipe WASM + DeepFace for real-time cognitive load tracking and engagement analysis.
Strict RAG pipeline grounding LLM dietary recommendations in verified Ayurvedic database — zero hallucinations. InceptionV3 food recognition for calorie tracking. Full system: patient app, doctor portal, video consultations, marketplace.
Sports biomechanics platform with computer vision and pose estimation for real-time form correction. Tracks joint angles, counts reps, and analyzes movement efficiency. Sports social feed + coaching management + AI diet planning.
Mean-variance portfolio optimization engine for Indian equities. Implements Markowitz Modern Portfolio Theory with efficient frontier computation, Sharpe ratio maximization, and real-time NSE data integration for asset allocation decision-making.
ML-powered fantasy cricket team predictor analyzing player form, pitch conditions, head-to-head stats, and weather data. XGBoost ensemble generates optimal Dream11 lineups with player-impact scoring.
End-to-end options pricing research combining VIX regime detection, GARCH volatility modeling, and LSTM deep learning to predict real-time bid/ask spreads with implied volatility surface reconstruction.
Authored paper analyzing temporal stability in loan defaults using Hazard Structure Drift Index (HSDI). Demonstrated how Random Survival Forests outperform Cox models in detecting non-linear risk shifts across shifting economic environments. Research funded by university seed grant.
75% Merit Scholarship. Focus on AI/ML, Data Structures, Algorithm Design. Active in hackathons and open-source.
Deep focus on statistics, machine learning theory, and data-driven decision making from India's top institute.