Hao-Jun Michael Shi is a Research Scientist at Meta since August 2021, focusing on scalable and distributed training optimization methods for ranking/recommendation, generative AI, and content understanding applications. Prior to this role, Hao-Jun was a Research Intern at Facebook from January 2019 to May 2019, where the development of the quotient-remainder hashing trick for embedding compression and the investigation of multi-pass training for deep learning-based recommendation systems occurred. Hao-Jun also served as a Research Assistant at Northwestern University from April 2017 to August 2021, conducting research on second-order optimization algorithms and developing the PyTorch-LBFGS module. Educational qualifications include a PhD in Industrial Engineering and Management Sciences from Northwestern University (2021), an MS from the same institution (2017), and a BS in Applied Mathematics from the University of California, Los Angeles (2016).
This person is not in the org chart
This person is not in any teams
This person is not in any offices