Bayrem Benkhoud

✅ Project: Company Knowledge Assistant (RAG + Agent) 🧩 Description: A smart assistant that helps employees query internal documents (project specs, HR policies, meeting notes, etc.) using natural language. The system uses Retrieval-Augmented Generation and LangChain agents to reason, search, and respond based on company documents. 🔧 Features: Upload company documents (PDF, DOCX, TXT, Markdown, etc.) Semantic search using vector database (RAG) Conversational UI with memory and context LangChain agent to answer complex queries with reasoning User roles: Admin (manage documents), Employee (chat assistant) Real-time web UI with Chainlit 📦 Backend (Python, LangChain, FastAPI): Document ingestion pipeline: Extract text from PDFs, DOCX, images (OCR) Chunking & metadata tagging Embedding with OpenAI, Cohere, or Gemini Store in vector DB (Chroma / FAISS / Weaviate / Qdrant) RAG setup: LangChain Retriever + ConversationalRetrievalChain Optionally ensemble retriever for multimodal 🎨 Frontend (Chainlit): Company-branded chatbot interface Role-based interaction: show admin upload panel, employee chat Message history, citations, source document highlights