Chongxiao Cao
Staff Machine Learning Engineer - Deep Learning Training Core at Snap Inc.
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolChongxiao Cao is a staff machine learning engineer with 12 years of experience building high-performance distributed systems for deep learning training and HPC. He has led multi-GPU training and data-loading infrastructure at Uber’s Michelangelo and now drives deep learning training core work at Snap, focusing on throughput, model scaling, and resource optimization. His PhD under Jack Dongarra anchors expertise in HPC, fault tolerance, and CPU–GPU heterogeneous computing, and he brings production-grade MPI and RDMA experience from Intel’s work on next-generation supercomputers. An active open-source maintainer and top contributor to projects like Horovod and Petastorm, he also contributed critical RMA and CH4 improvements to the flagship MPICH repository—bridging research rigor with systems-first engineering. Based in Seattle, he combines academic depth with pragmatic platform-building to accelerate large-scale model training.
12 years of coding experience
8 years of employment as a software developer
Master System Engineering, Master System Engineering at Xi'an Jiaotong UniversityPhD Minor Computational Science, PhD Minor Computational Science at University of Tennessee, Knoxville
Github Skills (5)
c1710
mpi10
message-passing10
c1110
fortran7
Programming languages (2)
Github contributions (5)
pmodels/mpich
Jun 2017 - Jun 2021
Official MPICH Repository
Role in this project:
Back-end Developer
Back-end DeveloperContributions:8 reviews, 120 commits, 34 PRs in 4 years 1 month
Contributions summary:Chongxiao contributed to the MPICH repository, focusing on improvements to the CH4 communication layer. Their commits primarily revolved around enhancing the Remote Memory Access (RMA) functionality, including implementing accumulate ordering for improved performance and stability. They also made several bug fixes and optimizations related to the handling of communicators and window synchronization within the CH4 architecture. The user's contributions directly enhance the capabilities of the Message Passing Interface implementation.
fortranhpcmpic
shawnccx/mpich
Jun 2017 - Aug 2020
Official MPICH Repository
Contributions:69 pushes, 27 branches in 3 years 2 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.