Hendrik Greving
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolHendrik Greving is a compiler and JIT systems engineer with 7+ years of professional experience and a long track record working on LLVM backends, software pipelining, and compiler-driven performance engineering. He has driven compiler efforts at Google (TPU LLVM backend), advanced Tesla's Dojo compiler, and now contributes to ML infrastructure as a Member of Technical Staff at OpenAI. His background spans system-level binary translation, custom GCC backends at Intel Labs, and GPU shader compilers for Samsung, giving him deep expertise across toolchains and hardware-aware optimization. Hendrik also contributes to notable open-source projects like DynamoRIO, improving Linux memory handling and low-level syscall performance. Based in San Francisco and trained at RWTH Aachen, he brings a rare combination of production-grade compiler implementation experience and recent focus on machine learning systems. A less obvious strength is his consistent work on the plumbing of runtimes and loaders—areas that materially boost performance but often sit below the visibility of higher-level ML stacks.
7 years of coding experience
19 years of employment as a software developer
Master thesis University of Technology "Analysis of HighDiplomingenieur, Information and Communication Technology, Diplomingenieur, Information and Communication Technology at RWTH Aachen University
Github Skills (6)
c1710
linux10
assembly10
instrumentation10
c1110
system-programming10
Programming languages (4)
Github contributions (5)
DynamoRIO/dynamorio
Oct 2018 - Nov 2019
Dynamic Instrumentation Tool Platform
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 review, 427 commits, 425 PRs in 1 year 1 month
Contributions summary:Hendrik's commits focused on enhancing the DynamoRIO dynamic instrumentation tool platform, specifically within the Linux environment. The contributions primarily involved modifications to memory management, with changes to how the library name is read in non-static Linux environments. The user was also responsible for improving the performance by implementing the new system calls ppoll, pselect, and epoll_pwait as well as fixing issues with the Linux loader.
instrumentationbinary-analysisprofilingwindowslinux
cutecutecat/dynamorio
Apr 2019 - Apr 2019
Dynamic Instrumentation Tool Platform
Contributions:1 push in 1 day
instrumentationdynamic-instrumentation
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.