Samrat Chanda

I'm an AI Engineer with 2+ years of experience building production LLM and RAG systems. At Accenture, I contribute to GenWizard, an internal Generative AI platform that automates authoring of technical documentation — I own the document ingestion and chunking pipeline, generate dense embeddings via Azure OpenAI, implement retrieval with cross-encoder re-ranking, and monitor pipeline quality using LangSmith and RAGAS. This work has helped cut manual document-authoring effort by up to 60% for delivery teams. Outside of my day job, I build things that push past tutorial-level GenAI. My latest project, Career Intelligence Copilot, is a fully Dockerized multi-agent career assistant — a 7-node LangGraph state machine with human-in-the-loop approval gates, a hybrid FAISS + BM25 RAG pipeline, a 12-tool registry (including custom MCP servers), and an 8/8 passing Playwright E2E test suite. I've also built a real-time computer vision pipeline (YOLOv8, Deep SORT, homography mapping) for sports performance analytics, processing 10K+ frames per match. My core stack is LangChain, LangGraph, Azure OpenAI, FAISS/vector search, Docker, and FastAPI, backed by a strong foundation in Python and classical ML/computer vision. I'm looking for a fully remote AI Engineer or GenAI Engineer role where I can own meaningful pieces of a production RAG or agentic system, ideally with a team that ships fast and cares about evaluation rigor, not just demos.