Description
Summary I am looking for an AI engineer to build a production-grade document intelligence system: a RAG pipeline and multi-agent layer that ingests my business documents, extracts structured data, answers grounded questions with citations, and exports clean reports. This is not a demo chatbot. We need something reliable, tested, and built to actually run in production. High volume of complex documents — contracts, filings, technical specs, and reports — and most of that work is still manual. I want a system grounded in our real content that can read these documents, pull out the fields that matter, answer questions accurately, and hand back structured output to me. Please suggest stack as per your convenience. Solution should be expandable in future. Send: - Two or three demos or live systems you've built (RAG, agents, or document automation) - A short note on how you'd approach v1 — the architecture, how you'd keep answers grounded, and how you'd measure quality - Your read on when an AI agent makes sense versus a plain deterministic workflow. - What do you think of using PydanticAI and langgraph in combination?