Danyang Zhuo

Danyang Zhuo is an Assistant Professor of Computer Science at Duke University, affiliated with the Department of Computer Science. His research spans datacenter networking and GPU systems, focusing on areas such as RDMA, collective communication, application networks, training, inference, and multiplexing. He has received several awards, including the NSF CAREER Award and multiple Best Paper Awards at top conferences. Zhuo previously served as a postdoctoral researcher at UC Berkeley and earned his Ph.D. from the University of Washington. His work has been recognized through numerous accolades and publications, and he is actively involved in organizing and reviewing at major conferences.

Danyang Zhuo is a researcher at Duke University specializing in Distributed Systems, Networking, and Operating Systems. His work focuses on optimizing distributed deep learning through automated parallelism, such as in Alpa and Ansor, and improving large-scale language model training with TeraPipe. He also investigates fairness in large language model serving and develops systems for multi-tenant inference, like Punica. His research addresses data center network challenges, including packet corruption and reliability, and explores low-overhead container networking with Slim. Zhuo's contributions span performance optimization, fault tolerance, and security in distributed systems, including service mesh sidecars and RDMA resource management. His work emphasizes scalable, efficient, and secure infrastructure for modern computing.