Graph Minds: Exploring Agentic Memory and Self-Improving AI

​We're excited to invite you to Graph Minds: Exploring Agentic Memory and Self-Improving AI, a special meetup co-hosted with the London Java Community at Microsoft Reactor!

​

​Join fellow graph enthusiasts, Neo4j developers and members of the Java community as we dive into cutting-edge topics like composable agentic memory and building self-improving AI agents with graph-based architectures.

​Let’s connect, learn, and explore how graphs are powering the next generation of intelligent systems.

​

​Session 1:

​Title: Building Self-Improving AI Agents with MCP and Graph-Based Memory

​Speaker: Opemipo Disu, Latitude

​

​Abstract:

​Autonomous AI agents are becoming more capable by the day, but how do we enable them to learn from experience and continuously evolve? In this talk, we'll explore how to build self-improving AI agents using the MCP (Memory-Cognition-Planning) architecture, integrated with Latitude to represent and reason over long-term memory. You’ll see how graphs can provide structured memory storage for agentic systems, support retrieval-augmented generation (GraphRAG), and facilitate feedback loops that help agents adapt to new environments and goals. Expect a live demo and actionable insights for building smarter, more context-aware AI systems.

​

​Session 2:

​Title: Composable Agentic Memory - from Local State to Institutional Knowledge

​Speaker: Andreas Kollegger, Neo4j