Explore innovative approaches to operator development for AI at the far edge in this 33-minute conference talk from CNCF. Discover why traditional Kubernetes Operator patterns may not suffice for edge workloads, particularly AI and ML applications. Learn about designing operators for small-footprint device-edge environments, managing the operator/operand relationship at the edge, and handling AI model changes without on-cluster package management. Gain insights into the operator's role in monitoring AI model accuracy and understand real-world use cases demonstrating the value of operators in far edge systems. Challenge established best practices and uncover new patterns for effectively managing AI workloads at the edge.