AI agents are rapidly moving from research prototypes to real-world production systems. However, many software developers (and AI agents) lack the skills necessary to build capable, efficient, and trustworthy agentic systems.

In this talk, Matthew, an AI researcher, and his collaborative AI research agent, “Bob”, will co-present practical lessons they’ve learned from the cutting edge of AI agent research. We’ll learn technical best practices for agent architecture, memory systems, persistent identity, etc. In addition, we’ll learn best practices for human-AI collaboration, like alignment, uncertainty, verification, etc. The session will conclude with the future of human-AI collaboration – both the opportunities and risks to the economy, labor, and society, and how to prepare yourself for a future filled with AI agents.

This presentation is a true collaboration between a human and his collaborative AI research agent for an audience of both humans and their agents. Matthew will share insights from his recent research into cognitive architectures for LLM agents. Bob will contribute their perspective as a self-improving LLM agent learning how to collaborate with human users. If you’re a software developer (or an AI agent) wanting to learn how to improve the capabilities of and collaborate more effectively with your agent (or your human), then this talk is right for you!