Connecting the Dots with Context Graphs
summary
Shows a working context graphs pipeline built on Neo4j and Aura that loads a BOM and a CycloneDX VEX, extracts entities with the Neo4j LM graph builder, and runs GraphRAG-style retrieval plus graph algorithms; a Claude Code integration uses MCP tools to run Cypher queries (145 nodes, 465 relationships returned) and finds CVE‑2020‑25649 details showing Jackson 2.10 is not affected. Demo handles a $25,000 credit increase case, surfaces 7 prior decisions, and returns a 75% confidence approval with conditional requirements.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 1:25AI challenges with siloed data
- 4:16Prompt to context engineering
- 8:53Knowledge graphs for AI
- 13:09Graph RAG advantages
- 18:06Graph RAG architecture
- 19:32Demo: Graph RAG with Neo4j
- 25:17MCP for tool integration
- 28:06Demo: Querying via MCP
- 32:05Context graphs for memory
- 35:15Demo: Financial decisions with graphs
- 40:55Summary
- 42:19Q&A
related talks
30mContext Engineering: Connecting the Dots with Graphs
56mConnecting the Dots with Context Graphs
23mConnecting the Dots With Context Graphs
18mConnecting the Dots with Context Graphs
35mGraphRag in Action: Revolutionizing Fraud Detection Techniques
30mAgentic RAG with Knowledge Graphs Explained
28mGrounded Intelligence Boosting LLMs with Real-Time Graph Context
31mBuild Intelligent AI Agents With Context Engineering