Hybrid RAG + Knowledge Graph
3 search methods × RRF fusion — RAGAS faithfulness 0.91 on 200 questions over corporate documentation (3K+ pages)
Architect & sole developer
2 мес · solo
- Python
- Qdrant
- Neo4j
- Sentence-Transformers
- RAGAS
- FastAPI
Knowledge OS — knowledge map (graph)
Expand the deep-dive ↓
What doesn't work
Vector search finds semantically similar documents but misses exact matches and entity relationships. Pure RAG over 3K+ page corporate documentation gives RAGAS faithfulness 0.67 — every third answer is incomplete or inaccurate.
Architectural approach
Three parallel search methods: Vector (Qdrant HNSW), BM25 (exact matches), Knowledge Graph (Neo4j, entity relationships). Results combined via Reciprocal Rank Fusion (RRF). Quality measured through RAGAS eval on 200 control questions.
Architect & sole developer
Designed and built from scratch: Qdrant HNSW setup, Neo4j knowledge graph with entity resolution, BM25 index, RRF fusion with weight tuning. Ran RAGAS eval on 200 control questions. Deployed as internal tool for corporate documentation.