I build production AI systems — LLM agents, RAG pipelines, and deep-learning models — end to end, from research prototype to deployed service. IIT Kanpur engineer turning applied ML into products people use.
- 🤖 Agentic AI & LLMs — design multi-agent systems and tool-using pipelines with LangChain, LangGraph, and the modern agent stack
- 🧠 Applied Deep Learning — CNNs, RNNs/transformers, and neural-network verification & repair research
- 🔎 RAG & retrieval — grounded, evaluated LLM applications with vector search and structured tool use
- ⚙️ Full-stack delivery — ship models behind React / Next.js frontends and Python (FastAPI / Streamlit) services, containerized for production
- 📈 ML for the real world — from customer-segmentation and forecasting to computer-vision challenges (ISRO Chandrayaan moon mapping)
Languages
AI / ML
Web & Infra
| Project | What it is | Stack |
|---|---|---|
| Agentic-AI | Multi-agent LLM system for autonomous, tool-using workflows | Python · LangChain |
| LangGraph Agentic AI | Graph-structured agent orchestration with LangGraph | Python · LangGraph |
| NNRepair (CSE-Research) | Constraint-based repair of neural-network classifiers via fault localization + solving | Python |
| Neural Style Transfer | CNN-based artistic style transfer, from research notebook to deployable app | TensorFlow · Python |
| MBTI Personality Prediction | NLP model classifying writing into 16 MBTI types | Python · NLP |
| Tweet Emotion Recognition | RNN multi-class classifier over 6 emotions | TensorFlow · Python |
| ISRO Moon Mapping | Super-resolution mapping for Chandrayaan-2 imagery (Inter-IIT Tech Meet) | Python · Computer Vision |
| Blockchain Aadhaar Voting | Decentralized voting system | React · Solidity |
Diving into the quantum realm with code — IIT Kanpur.