From Planning To Production-Ready RAG
summary
Demonstrates a deployable edge RAG stack built from the Open Edge Platform: modular AI libraries and AGI suites (image/video search, robotics), model servers OVMS and TGI, vector stores like Milvus, observability via OpenTelemetry, and Helm/docker deploys. Shows a multi‑agent pipeline that extracts entities, builds a Neo4j knowledge graph, runs graph retrieval strategies and auto‑constructs schema from CSVs (products/parts/suppliers). Claim: converting chunks to entities + relations multiplies retrievable answers (n → ~n^2 or more). The demo actually builds, resets, and queries the graph successfully.
AI-written summary · about
chapters (14)
- 0:00Session introduction
- 1:25Deploy AI models on edge
- 8:15Edge Platform architecture
- 16:44Edge AI use cases & demos
- 17:27Graph RAG introduction
- 20:18Relational to graph databases
- 23:04What makes graph RAG different
- 31:53Graph RAG data sources
- 33:38SEC financial data walkthrough
- 43:08Multi-agent knowledge graph
- 47:42Multi-agent system architecture
- 50:08Single-agent system demo
- 53:00Full multi-agent system demo
- 59:45Summary
related talks
21mHow RAG is Redefining AI Capabilities
16mOpenRAG: An open-source stack for RAG
24mScalable GenAI in Production
12mRevolutionizing Platform Engineering with AI Agents and GraphRAG
27mAdaptive Knowledge Graphs for Dynamic RAG Systems
31mAdvanced Enterprise RAG Systems
01h 17mGraph Database and Retrieval Augmented Generation
39mEnhancing RAG with Neo4j Knowledge Graph