Tommy H.

I build production RAG and AI agent systems for operators who need answers grounded in their own documents — not generic chatbots. Recent work includes an insurance knowledge base with document ingestion, retrieval-augmented Q&A with citations, and a human review digest so nothing becomes decision-ready until someone checks it. Same pattern across Django/Python + Postgres platforms, MCP tool fleets, Hermes-style agent workflows, and CRM/ops automations. What I typically deliver (hourly): - RAG pipelines over PDFs / knowledge bases (chunking, retrieval, grounding, eval/review loops) - Lightweight agents and connectors on your existing stack - CRM and back-office automations that cut admin hours Hourly only. No Microsoft Copilot Studio / Azure AI Foundry stack. Clear requirements, working artifacts early, no Loom until we’re actually talking.