AI-ready control infrastructure for cyclotron systems using GPU-accelerated Python GUIs and LabVIEW over ZeroMQ

We present a modular, AI-ready control and monitoring infrastructure developed for the 76-inch isochronous cyclotron at the Crocker Nuclear Laboratory, University of California, Davis. The system combines a GPU-accelerated Python GUI engine on a high-performance Linux workstation with a LabVIEW-based supervisory platform for real-time control and data acquisition. Communication between platforms is handled via ZeroMQ, enabling low-latency, asynchronous data exchange. Benchmark results show end-to-end response times below 10 ms with minimal jitter, supporting real-time visualization and interactive feedback. Designed to separate deterministic control from high-level logic and user interaction, this architecture offers robust performance, scalability, and extensibility. It lays the groundwork for future integration of AI-based optimization, autonomous control, and predictive diagnostics in cyclotron operations.