Build Daily

Tinley Park · August 23, 2026
Neo4jAug 14, 2026

Building the Future of Agent Memory with Vasilije Markovic

What it covers

The speaker discusses their company's work on agent memory systems using Neo4j at Graph Gathering, exploring how agents connect to enterprise data and deploy in swarms. They also reflect on how AI tools like cloud code are democratizing development and predict a future where personal knowledge is distributed through avatars.

The outline

Tap a timestamp to jump there in the video.

  1. Agent Memory at Graph Gathering

    The speaker introduces their company's focus on agent memory using Neo4j and the opportunity to discuss implementations with industry partners.

  2. Defining Agent Memory Use Cases

    They mapped out definitions for agent memory, exploring how agents connect to enterprise data and deploy in swarms.

  3. Learning from Industry Experts

    The speaker highlights the value of Graph Gathering as a collaborative space with experts building with graphs for over a decade.

  4. Cloud Code and New Developers

    They discuss how cloud code enables non-developers to build complex systems using prompts and instructions in a terminal.

  5. Workshop with Cloud Code

    The speaker describes a recent workshop where participants built integrations using memory without leaving the terminal.

  6. Career Shift to AI Systems

    Reflecting on their past in data pipelines, they note how their old job is now non-existent due to AI and dynamic memory systems.

  7. Future of Personal Knowledge Avatars

    They predict a future where personal knowledge is distributed through avatars that can be rented out to enterprises globally.

Memory needs structure

Agent memory becomes useful when it connects to enterprise data rather than sitting in isolated logs. For solo builders, that points toward graph-based systems as a practical way to give agents context without overcomplicating the stack.