CaMLSys Lab

The Cambridge Machine Learning Systems Lab (CaMLSys) at the University of Cambridge investigates foundational challenges at the intersection of machine learning and computational systems. Our research spans two main areas: (1) developing theoretically grounded machine learning methods with strong real-world applicability, emphasizing both performance and safety — as seen in our work on federated learning and AI safety; and (2) designing algorithms, system software, and hardware that treat machine learning as a first-class workload, enabling major advances in efficiency, scalability, and deployment. One example is Photon, our system for decentralized training of large language models