Developing Graph RAG Part II

THIS WILL BE A HYBRID TALK. The online link will be provided closer to the talk.

In my earlier session, I built a foundation for an AI-driven career development tool using Role Ontology Graphs and Retrieval-Augmented Generation (Graph RAG). This follow-up dives deeper into the engineering behind simplifying that stack using Neo4j’s native capabilities.

We’ll explore how Neo4j now supports integrated vector search via native indexes, removing the need for external vector databases like FAISS. You’ll see how to build a similarity search directly on top of your role ontology to enable personalized, context-aware career guidance using just Cypher and GDS.

I’ll also demonstrate how to rapidly prototype and extend graphs using tools like arrows.app and Neo4j Graph Builder app—enabling interactive design, skill/role modeling, and testing schema evolutions with minimal overhead.

To close, I’ll introduce Neo4j MCP (Model Context Protocol) server and discuss how it enables modular, scalable backends for agentic systems.
This session is built for engineers, architects, and applied AI developers looking to reduce operational complexity and explore tighter integrations between knowledge graphs, semantic search, and LLM agents.

Big Data Bellevue Meetup was created by Intelius and takes place in downtown Bellevue. Intelius provides the only centralized service for delivering comprehensive information about people, places, organizations, and their connection to each other. Our state-of-the-art big data technology platform is utilized across a wide range of industries to implement specific solutions.

On the third Thursday of each month, we invite an industry leader in Big Data to give a presentation followed by a lively discussion on big data technology and its impact on business world. Past speakers include researchers from the University of Washington, as well as senior members of various companies, such as Microsoft, Amazon, eBay, IBM, MapR and inome.

Microsoft graciously provides free refreshments!