Designed and built a production-ready deployment pipeline for RAG (Retrieval-Augmented Generation) chatbots, enabling rapid deployment across multiple client environments. Key highlights: • Built a reusable pipeline to ingest, chunk, and embed documents into a vector store for retrieval. • Supports multiple LLM backends (OpenAI, Anthropic) with configurable prompt templates. • Automated deployment workflow using CI/CD pipelines for zero-downtime releases. • Modular architecture allows swapping embedding models, vector DBs, and LLMs independently. • Used internally at FactSet to power document Q&A and internal knowledge assistants. Tech stack: Python, Node.js, LangChain, OpenAI, Pinecone/pgvector, AWS EC2, Docker