• Engineered a dynamic RAG system using LangGraph that intelligently routes queries between vector databases (ChromaDB), real-time web search (Tavily API), and foundational LLMs (Groq) based on semantic analysis. • Developed a production-ready Streamlit interface with document ingestion (PDF/TXT), vector store management, and transparent source attribution, enabling user overrides and session-based evaluation. • Architected modular components for retrieval fallbacks and error handling, containerized with Docker for consistent deployment, and integrated LangSmith for performance tracing.