PC Troubleshooting RAG

A document-grounded RAG assistant that answers PC hardware troubleshooting questions using a real staff training guide.

PythonClaude APIScikit-learnSentence Transformers
SYS.ACTIVE
RAGKNOWLEDGE CORE

Building a useful system.

This Python RAG system retrieves relevant sections from a PC diagnostics guide and asks Claude to answer only from that evidence with citations. It compares deterministic TF-IDF retrieval against sentence-embedding search: semantic retrieval improved the fixed evaluation set from 8/10 to 10/10 correct results.

View GitHub ↗← Back to Projects