EDS AI for Science and Operations

The AI for Science and Operations (ASO) department, situated within the LCLS Experiment Data Systems Division, is dedicated to the practical integration of artificial intelligence into the lifecycle of high-repetition-rate X-ray science. Our primary focus is the development and maintenance of robust AI for Operations, providing operator-assistance and automation tools that ensure facility reliability and machine protection during 24/7 production. By moving beyond isolated demonstrations, we build hardened infrastructure capable of managing the extreme data throughput of LCLS-II and LCLS-II-HE, ensuring that AI components function as predictable, mission-critical assets for the laboratory.

In parallel, the department leads AI for Science as a core R&D initiative, exploring agentic frameworks to assist researchers in experimental steering and autonomous discovery. This work aims to help users navigate complex, high-dimensional parameter spaces in real-time, maximizing the scientific yield of limited beamtime. Underpinning all our efforts is a commitment to safety and validation; we establish standardized patterns for monitoring and evaluation to ensure that all deployed models—from edge-based data reduction to autonomous steering—remain verifiable and aligned with the rigorous requirements of the scientific community.