Composable RAG Pipelines With OPEA Highlighting GraphRAG
summary
Open Platform for Enterprise AI (OPEA) provides composable microservices that assemble RAG pipelines using vector DBs (Milvus), an embedding microservice and a Neo4j-backed GraphRAG with Neo4j knowledge graph microservice, plus an agent supervisor that picks tools. Demo uses the EDGAR SEC filings (EDGAR SEC dataset), runs hybrid vector + graph queries, returns sensible results in ~45s, scales as microservices, and exposes a remaining hallucination case (misclassified Nike) while the project grows from 14 to 47 partners.
AI-written summary · about
chapters (7)
- 0:00Introduction and context
- 3:00Why open source for GenAI
- 5:28OPEA and composable RAG pipelines
- 9:05Introducing GraphRAG concept
- 12:58GraphRAG demo with Neo4j
- 19:28Microsoft GraphRAG demo
- 23:01Community and next steps
related talks
38mLeveraging the OPEA Platform and GraphRAG Architectures To Drive Cl...
28mAI Pipelines With OPEA: Best Practices for Cloud Native ML Operati...
40mRAG Pipeline Evaluation
35mGraphRag in Action: Revolutionizing Fraud Detection Techniques
38mIntroducing Open Platform for Enterprise AI
01h 36mTutorial: From Planni...
27mAdaptive Knowledge Graphs for Dynamic RAG Systems
16mOpenRAG: An open-source stack for RAG