← back
Intro to GraphRAG — Zach Blumenfeld
Takeaway
Even a simple knowledge graph plus LangGraph agent gives you accurate, explainable retrieval that beats raw vector search for structured-domain questions.
Summary
- Hands-on Neo4j workshop building a skills-and-employee knowledge graph for talent search, staffing and team formation use cases.
- Three modules: graph basics with Cypher queries, unstructured-data entity extraction, and a simple LangGraph agent with retrieval tools.
- Architecture pattern: KG sits between data sources (structured+unstructured) and the agent, exposing domain logic via schema so retrieval is more controllable and explainable than one-shot vector search.
- Argues KG matters more in agentic workflows because prompts get decomposed into multiple sub-queries and graph traversal complements vector retrieval.
graphragneo4jlanggraph
Original description
Learn the foundations of GraphRAG, starting with knowledge graph construction and then common retrieval patterns. --- GraphRAG has gone from nice-to-have to essential as AI solutions have increased in sophistication. This workshop will get you started, answering: - what is GraphRAG, and when do I need it? - what's the best way to construct a knowledge graph? - how do I combine unstructured and structured data? - how do I retrieve the right information? About Zach Blumenfeld Zach Blumenfeld is a Data Science Product Specialist at Neo4j who helps empower the market with Neo4j’s industry-leading graph data science capabilities. He has first-hand experience with various modern-day DS/ML challenges, including criminal fraud detection, identity resolution, and recommendation systems. Having served in both data science and software developer capacities, Zach has applied graph computing to law enforcement and government entities in support of missions that counter drug trafficking, human smuggling, money laundering, and child exploitation. He has led the development and deployment of full-stack graph systems designed to facilitate broad data science and operational requirements. Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter