Yi Cheng
About Yi Cheng
Yi Cheng is a software engineer with extensive experience in developing scalable systems and contributing to open-source projects. He has worked for notable companies like Tencent, Baidu, Facebook, and TigerGraph, and is currently employed at Anyscale, where he focuses on enhancing the Ray software framework.
Work at Anyscale
Currently, Yi Cheng serves as a Software Engineer at Anyscale, a position he has held since 2021. He operates in the San Francisco Bay Area in a hybrid work environment. In this role, he focuses on enhancing the reliability and scalability of the open-source software Ray, contributing significantly to its development. Yi has led initiatives to improve Ray Serve's availability and has been instrumental in building a workflow engine on top of Ray.
Previous Experience
Prior to his current role, Yi Cheng worked at several prominent technology companies. He was a Software Engineer at TigerGraph from 2019 to 2021, and before that, he spent three years at Baidu, Inc. from 2011 to 2014. He also held positions at Facebook, where he worked as a Software Engineer from 2016 to 2019 and completed an internship in 2015. Additionally, he gained early experience as a Software Engineer Intern at Tencent in 2010.
Education and Expertise
Yi Cheng holds a Master of Computational Data Science from Carnegie Mellon University, where he studied from 2014 to 2015. He also earned a Bachelor's degree in Computer Science from Sun Yat-sen University, completing his studies there from 2007 to 2011. His educational background has equipped him with a strong foundation in data science and software engineering.
Research and Publications
Yi Cheng has made significant contributions to the field of software engineering through research and publications. He co-authored the paper 'ExoFlow: A Universal Workflow System for Exactly-Once DAGs,' which was presented at the USENIX OSDI 2023 conference. Additionally, he published a paper titled 'Loading Llama-2 70b 20x faster with Anyscale Endpoints,' showcasing his work on performance optimization.
Contributions to Open Source Software
Since 2021, Yi Cheng has been a major contributor to the open-source software Ray. His contributions include scaling Ray clusters from 200 to over 4,000 nodes, which significantly reduced network usage and CPU utilization. He has also assisted OpenAI in adopting Ray as their training platform by enhancing its scalability and scheduling performance.