Zaid Zaim

Zaid Zaim is an AI developer advocate at Neo4j and a Microsoft MVP who helps developers build AI agents grounded in organizational relationships, persistent memory, and explicit rules. His career began in immersive computing, where he used mixed reality to reconstruct endangered Syrian heritage, before moving into graph databases and AI-agent infrastructure.

Zaim grew up in Damascus and moved with his family to Germany in 2013 after war forced them to leave Syria. At Berlin’s ReDI School of Digital Integration, volunteer mentors and hackathons introduced him to Microsoft HoloLens and immersive software development.

Working with developers and artists, he created a mixed-reality reconstruction of Palmyra, the ancient Syrian city whose monuments were damaged during the conflict. Built with HoloLens, Unity, and Microsoft’s Mixed Reality Toolkit, the project recreated historic structures and led to a TEDx Munich appearance.

By 2022, he worked as a mixed-reality catalyst at Hololux, developing HoloLens applications for remote maintenance and immersive collaboration. He also worked with the consultancy ignore gravity, studied at CODE University, and developed public mixed-reality and virtual-reality projects. At Microsoft Build 2024, he helped developers explore immersive copilots built with Microsoft Mesh and Azure OpenAI.

Building agents with durable context

At Neo4j, Zaim works where graph databases meet agent frameworks, with particular attention to the Microsoft developer ecosystem. His technical priorities include:

• Context graphs for decision-aware agents: Connect people, departments, documents, policies, and previous actions so agents can understand why a particular organizational decision makes sense.

• Three complementary forms of agent memory: Combine short-term conversation history, long-term knowledge of people and organizations, and reasoning memory that captures relevant policies and rules. His writing on context graphs and AI memory explores how these relationships preserve context beyond conventional semantic retrieval.

• Microsoft Agent Framework and Neo4j integration: His guide to graph-aware Microsoft agents outlines four implementation patterns: automatic context providers, direct Cypher-query tools, shared Model Context Protocol servers, and HTTP interfaces for restricted environments.

• Agentic GraphRAG and Text-to-Cypher: Translate natural-language requests into graph queries, then traverse explicit organizational relationships to retrieve relevant knowledge. At AI Engineer Europe 2026, Zaim explained these foundations before fellow Neo4j presenter Andreas Kollegger introduced a separate agent-decision framework.

Zaim has also demonstrated these integrations at Microsoft AI Tour, including GitHub Copilot Text-to-Cypher tools and Neo4j context providers for Microsoft Agent Framework.