Company-website-chatbot-using-RAG-pipeline-and-LLM-intelligence

This project is an AI-powered company website chatbot built using a Retrieval-Augmented Generation (RAG) pipeline and Large Language Models (LLMs). The chatbot retrieves relevant information from a knowledge base before generating a response, helping provide context-aware and grounded answers instead of relying only on the language model's internal knowledge. The system implements a complete RAG-based conversational pipeline that combines document retrieval, hybrid search, reranking, conversation memory, and LLM-based response generation. It is designed to answer questions using information retrieved from company-related documents while maintaining conversational context. Key features include semantic search using embeddings, BM25 keyword-based retrieval, hybrid search combining semantic and lexical retrieval, Reciprocal Rank Fusion (RRF), cross-encoder reranking, context-aware prompt construction, conversation memory, query normalization and routing, document cleaning and chunking, metadata processing, and LLM-based response generation. The backend is built with FastAPI and supports local Ollama, Groq LLM integration, Google GenAI integration, and ChromaDB vector storage.