AI agents are rapidly evolving from experimental demos and are today used in biological research, data automation, and data science. But today’s agent ecosystem is fragmented: each agent comes with its own execution environment, tooling requirements, logging, and pricing. This makes it difficult for organizations to integrate, evaluate, and switch between agents.

In this session, we’ll explore the concept of a Control Plane for Agents: a unified layer that standardizes how AI agents are accessed, executed, monitored, and managed. Inspired by what platforms like OpenRouter did for LLMs, the goal is to discover, evaluate, and switch between agents through a single API.

Participants will learn about the emerging landscape of open-source agents, the challenges involved in running them in production, and how a unified control plane can simplify integration while enabling flexibility. 

What Participants Can Expect
•    A practical overview of the rapidly growing AI agent ecosystem
•    Discussion of the key challenges when integrating agents
•    Introduction to the concept of a Control Plane for Agents
•    Examples of how a single layer can enable flexible agent integration
•    Open discussion with builders and researchers working on agent technologies

This session is designed for:
•    AI practitioners in biology evaluating or using agent-based systems
•    Researchers and data scientists exploring agent workflows
•    Leaders interested in the future of agent-native software
Anyone curious about the next generation of AI software infrastructure is welcome to join.

Registration: This event is free to attend. Please email [email protected] to request an invitation.