Ming Liu
Senior Staff Software Engineer at Google
Greater Seattle Area United States
Summary
🤩
Rockstar🎓
Top SchoolMing Liu is a Sr. Staff Software Engineer based in the Greater Seattle Area with seven years of experience architecting cloud-scale AI accelerators and full-stack system solutions. He blends hands-on hardware design and software toolchain expertise, having led custom AI accelerator projects from architecture through delivery while also managing teams and driving research. At Google and Microsoft he contributed deep systems optimizations—e.g., improving TensorFlow Serving model warm-up monitoring and enhancing LLVM ThinLTO bitcode handling—to boost performance and reliability at scale. His background includes advanced EECS study at MIT (ABD) and a B.A.Sc. in Computer Engineering from University of Toronto, reflecting strong academic rigor underpinning pragmatic engineering. Notably, he moves fluidly between low-level compiler and hardware concerns and high-level production monitoring, making him effective at closing the loop from silicon to serving. Colleagues rely on him for complex cross-disciplinary problem solving and measurable system performance wins.
7 years of coding experience
9 years of employment as a software developer
PhD (Did Not Complete - ABD), PhD (Did Not Complete - ABD) at Massachusetts Institute of TechnologyBachelor of Applied Science (B.A.Sc.), Computer Engineering, Bachelor of Applied Science (B.A.Sc.), Computer Engineering at University of Toronto
Chinese, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (13)
bitcode10
machine-learning10
c-language10
tensorflow10
cprogramming-language10
ml10
llvm10
batching9
performance-optimization9
compiler-optimization9
batch9
deep-learning8
python8
Programming languages (4)
Github contributions (5)
llvm/llvm-project
Sep 2021 - May 2026
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:509 reviews, 310 PRs, 532 pushes in 4 years 8 months
Contributions summary:Ming contributed to the LLVM project, focusing on ThinLTO (Thin Link-Time Optimization) and the related bitcode handling. Their work included enhancing the generation of import type information within bitcode, addressing symbol handling, and optimizing memory usage during LTO indexing. These changes aimed to improve the efficiency and correctness of the compilation process, particularly for distributed ThinLTO builds. The user also worked on updating and maintaining test cases to reflect these changes.
compilerllvmtoolchain
tensorflow/serving
Jan 2019 - Aug 2021
A flexible, high-performance serving system for machine learning models
Role in this project:MLOps Engineer
MLOps EngineerContributions:10 releases, 79 commits, 44 PRs in 2 years 7 months
Contributions summary:Ming's primary contributions focused on improving the monitoring and performance of model warm-up procedures within the TensorFlow Serving system. They added monitoring metrics and used a sampler to record warm-up latency distributions, allowing for better tracking of performance. The user also made changes to protobuf handling, specifically cleaning up usage of the protobuf_archive and related dependencies. Furthermore, the user made code changes to support improved tensor serialization and added support for large batch splitting in the batching session.
machine-learning-modelsmachine-learningdeep-learningdeep-neural-networkspython
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.