Local RAG System
AI Engineer & Backend Developer
Government Ordinances
A question-answering system for municipal ordinances and memos. It runs on local models only, so government text never leaves the machine.
Built for a client. The source is not public, the full story is right here.
Overview
Ordinances and memos are the documents people most need to search and the documents government cannot paste into cloud chatbots. This system answers questions over that corpus, and every model call stays on local hardware.
Architecture
A Python pipeline chunks documents, embeds them, and stores them in a local vector index. At query time it retrieves by similarity and generates answers with Ollama serving Llama 3 8B or Mistral. Retrieval and generation are separate layers, so either model can be swapped without touching the other.
Outcomes
- Answers grounded in the source documents
- Fully offline, no data leaves the machine
- Swap between Llama 3 8B and Mistral behind one interface
- The pattern fed into PASADA's document pipeline