Events for March 11, 2026

Abstract: Agentic AI systems represent a shift from stateless model inference to long-running, autonomous, and distributed intelligent entities that reason, coordinate, and act within complex environments. This session frames Agentic AI as a systems design problem and draws on established principles from distributed systems to introduce reusable architectural patterns for building reliable and governable agentic systems. Key patterns—including planner–executor separation, consensus-based decision making, event-driven agents, and shared memory with eventual consistency—are examined through real-world system behaviors and failure modes. The discussion highlights how these patterns improve robustness, observability, and human oversight while mitigating risks such as hallucination propagation and uncontrolled autonomy. The session concludes by outlining open research challenges in verification, consistency, and testing of emergent agent behavior. Speaker(s): Subbarao Pydikondala, Room: Third Floor, ECNT 312 Conference Room (B_ECNT), Bldg: 422 UCB, 1111 Engineering Drive, Boulder, CO , Colorado, United States, 80309-0422