AI Chatbot Development (RAG)

About this Gig

I build chatbots that answer questions strictly from your own documents — using Retrieval-Augmented Generation (RAG), not generic AI guesswork. The chatbot retrieves relevant information from your actual content before generating a response, so it never invents answers. What I've built and tested: - A restaurant chatbot answering menu, hours, and policy questions purely from source documents - Verified fallback behavior — the bot correctly refuses to answer when information isn't in your documents, rather than hallucinating a wrong answer - Document chunking, vector embeddings, and similarity search pipeline (Chroma DB) I'm upfront about scope: I work with Chroma DB as the vector store (not Pinecone or Weaviate specifically), and builds are tested locally/via secure tunnel rather than deployed to a permanent production server — happy to discuss your specific hosting needs before we start.

Requirements

To begin, I'll need: 1. The source documents your chatbot should answer from (PDFs, text files, FAQs, policies, etc.) 2. A sense of your typical customer questions, so I can test edge cases properly 3. Where you'd like the chatbot deployed or embedded (website, WhatsApp, etc.) — this affects scope 4. Any specific tone/persona you want the bot to have when responding

Related Tags