Daytona AI Researchers - Berkeley, October 2026
- WHO IT IS FOR: AI researchers, developers, and academics interested in the future of Reinforcement Learning and agentic workflows.
- WHAT IT IS ABOUT: Exploring the evolution of AI agents, moving from stateless episodes to long-horizon, sessionful environments and scalable primitives.
- FORMAT: A series of technical talks followed by a structured networking session with food and beverages.
Join us in Berkeley for the inaugural Daytona AI Researchers meetup, a dedicated gathering designed for the academic and technical community at UC Berkeley and beyond. This event dives deep into the cutting edge of artificial intelligence, specifically focusing on how modern AI agents are outgrowing traditional reinforcement learning abstractions. We will explore the shift from short, stateless episodes to complex, long-horizon tasks that require persistent environment states and advanced branching capabilities.
Our featured technical session, led by Muhammad Annas Hashmi from Daytona, will examine the critical need for new primitives like cheap VM forking and mid-run snapshots to support the next wave of agentic RL. Following the presentations, attendees will have the opportunity to engage in meaningful discussions and professional networking over food and drinks. Whether you are researching long-horizon benchmarks or speculative RL algorithms, this is the place to connect with peers driving the future of AI.