Designed an end-to-end Retrieval-Augmented Generation (RAG) customer support chatbot that gives context-aware, document-grounded answers instead of generic LLM responses.
- Orchestrated retrieval and generation with LangChain and OpenAI APIs.
- Ingested and vectorized company documents in PostgreSQL with pgvector for efficient semantic search.
- Deployed a containerized React + FastAPI application on AWS using Docker for scalable cloud hosting.