Built a production-grade AI customer support chatbot for ShopEase using a hybrid RAG + GraphRAG pipeline. The system combines semantic search with knowledge graph traversal to retrieve the most relevant responses from a 1,350-document knowledge base. Features include multi-hop graph traversal, dynamic graph learning from conversations, interactive knowledge graph visualization, conversation memory, and SQLite interaction logging. Flexible LLM support — works with cloud APIs (Groq, OpenAI) or fully local models via Ollama.