Xiaozhu Meng
Software Engineer at Meta
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolXiaozhu Meng is a software engineer and GPU performance tooling specialist with 14 years of systems and research experience, now working at Meta from San Jose. He has led development of GPU performance analysis across AMD, Intel, and NVIDIA stacks and built production-grade capabilities that surfaced real-world bottlenecks in frameworks like PyTorch. An expert in machine-code analysis and binary instrumentation, Xiaozhu contributed core x86-64 and PowerPC instruction semantics to the widely used DyninstAPI and helped scale tooling used in Linux distributions. His background blends academic rigor (PhD, UW–Madison) with industry impact at Amazon Lab126 and collaborations with vendors and DOE projects, making him equally comfortable in prototype research and shipping robust engineering. Notably, he has driven multi-architecture DWARF parsing and instrumentation improvements that enabled analysis of very large applications such as Firefox.
14 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at University of Wisconsin-MadisonBachelor's degree Computer Science and Technology, Bachelor's degree Computer Science and Technology at Huazhong University of Science and Technology
English, Chinese
Github Skills (10)
c-language10
architectures10
cprogramming-language10
binarydiff10
computer-architecture10
architecture10
instruction-set10
assemble9
assembler9
assembly9
Programming languages (6)
Github contributions (5)
dyninst/dyninst
May 2014 - Apr 2022
DyninstAPI: Tools for binary instrumentation, analysis, and modification.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:3 releases, 26 reviews, 495 commits in 8 years
Contributions summary:Xiaozhu primarily contributed to the development of x86-64 instruction semantics within the DyninstAPI, focusing on binary instrumentation, analysis, and modification. They added support for various powerpc instruction. Their contributions involved implementing instruction semantics and addressing issues related to memory access and control flow within this binary analysis tool. The commits highlight work related to supporting new instructions and ensuring correct behavior of the tool.
instrumentationbinary-analysisstackwalkermodificationbinary-instrumentation
mxz297/dyninst-tools
May 2019 - Jun 2021
A collection of command-line tools based on Dyninst
Contributions:74 commits, 61 pushes, 1 branch in 2 years 1 month
command-line-tool
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.