Stephen Herbein
Senior Systems Software Engineer at NVIDIA
Wilmington, North Carolina, United States
Summary
🤩
Rockstar🎓
Top SchoolStephen Herbein is a Senior Systems Software Engineer with 11 years of experience building high-performance, production-grade systems—currently at NVIDIA and with a Ph.D. in Computer Science from the University of Delaware. He specializes in HPC schedulers, resource managers, and system-level tooling, having developed hierarchical and I/O-aware scheduling that delivered multi-fold speedups and reduced scheduling complexity during research at LLNL. An active open-source contributor, he has improved Python bindings for the ZFP compression library and strengthened cross-platform package builds in Spack, reflecting deep expertise in build systems and scientific software packaging. Based in Wilmington, NC, he blends rigorous research with hands-on engineering to ship stable, performant infrastructure for large-scale compute environments. A non-obvious strength: he pairs low-level memory and error-handling improvements with usability-focused bindings, bridging C/C++ libraries to Python ecosystems for broader adoption.
11 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Computer Science, 4.00, Doctor of Philosophy (Ph.D.), Computer Science, 4.00 at University of Delaware
Github Skills (24)
python10
build-system10
numpy10
package-manager-tool10
cython10
compression10
package-manager10
typed-array9
multidimensional-arrays9
data-structure9
dynamic-array9
arrayobject9
scientific-computing9
parallel-arrays9
data-structures9
Programming languages (20)
Github contributions (5)
LLNL/zfp
Dec 2018 - Apr 2019
Compressed numerical arrays that support high-speed random access
Role in this project:
Back-end Developer
Back-end DeveloperContributions:41 commits, 1 PR, 30 comments in 3 months
Contributions summary:Stephen significantly contributed to the python bindings for the ZFP library, focusing on improving the usability and stability of the library. They added error checking for memory allocation and decompression, as well as implemented initial python bindings with support for NumPy arrays, including compression and decompression functions. The user also integrated test utilities and improved the build process by including CMake modules and providing documentation.
radiussspeedcompressiondata-vizeigen3
spack/spack
Nov 2015 - Oct 2021
A flexible package manager that supports multiple versions, configurations, platforms, and compilers.
Role in this project:Software Engineer (focused on package management and build tools)
Software Engineer (focused on package management and build tools)Contributions:6 reviews, 24 commits, 20 PRs in 6 years
Contributions summary:Stephen's contributions primarily involve modifying and improving the Spack package manager. They fixed bugs related to command-line flags, dependency resolution, and build processes for various packages. They also added support for new versions of dependencies like Flux prerequisites and updated build configurations to ensure compatibility across different operating systems and compilers. The user demonstrates expertise in package management principles by modifying build scripts and integrating external libraries within the Spack framework.
compilerspythonradiussplatformslinux
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.