AI Engineer at Meltwater. I build AI agents and LLM systems that work in production. Previously a Data Scientist at BNP Paribas and American Express · IIT Madras (B.Tech + M.Tech, Data Science)
- Agentic AI: production AI agents with LangGraph, Claude, and the Model Context Protocol (MCP)
- Context engineering: making agents see the right information on every turn, so they stay accurate, fast, and affordable
- LLM fine-tuning and evals: LoRA fine-tuning of open-source models, and LLM-as-judge evaluation
- Retrieval and document AI: RAG with ColBERT, HyDE, and reranking, plus OCR and document understanding
- Work on Mira, Meltwater's agentic AI assistant, offered to 27,000 customers including Microsoft
- Lead Synapse, a 4-person R&D team. I set the research direction and designed 30+ experiments across classic ML, embeddings, and LLMs
- Built an LLM trade-validation app at BNP Paribas, projected to drive €20M in annual profit
- Winner of 2 internal hackathons at Meltwater · Rank 2 of 225 in a national data science championship
| Project | What it is | Stack |
|---|---|---|
| Regulation RAG Assistant | Cited Q&A over EU banking regulation with ColBERT and HyDE | ColBERT, HyDE, Streamlit |
| EcoAI | Private local-LLM document chat with proposition-level retrieval | Ollama, ChromaDB, reranking |
| GuessWars | A transformer written from scratch that plays Hangman (63% win rate) | PyTorch, self-play |
Python LangGraph LangChain MCP Claude OpenAI PyTorch Hugging Face LoRA LangSmith RAG FastAPI PostgreSQL Redis Docker Kubernetes Terraform AWS Azure
LinkedIn · Portfolio coming soon