Steve S
Software Architect at NVIDIA
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolSteve S is a Software Architect in the San Francisco Bay Area with 11 years of experience designing software where physical systems meet computation, currently leading test automation architecture for NVIDIA’s Silicon Solutions Group. His background spans embedded firmware, hardware lab automation, and large-scale test infrastructure from roles at NVIDIA, Facebook, Lyft, Oculus and Microsoft, blending low-level device expertise with cloud-scale telemetry and CI pipelines. He’s a hands-on systems programmer—authoring high-performance C++ log readers and a columnar timeseries library—and has repeatedly improved developer workflows to catch issues earlier. An active contributor to high-profile open-source projects like Open MPI and NVIDIA’s NCCL repos, he’s fixed CUDA-related segfaults, eliminated potential deadlocks, and added performance-focused tests that strengthen multi-GPU communication stacks. Steve pairs advanced academic training in CS and EE with practical RF, sensor, and vision experience, uniquely enabling end-to-end solutions from hardware measurement to production software. Colleagues rely on him for pragmatic architectures that reduce pain points for developers while scaling complex hardware-software testbeds.
10 years of coding experience
22 years of employment as a software developer
Master of Science (M.S.) Electrical Engineering, Master of Science (M.S.) Electrical Engineering at University of WashingtonMaster’s Degree Computer Science, Master’s Degree Computer Science at Georgia Institute of TechnologyB.Sc. Electrical Engineering, B.Sc. Electrical Engineering at Université de SherbrookeCertificate Digital Signal Processing, Certificate Digital Signal Processing at UCSD - Extension
English, French
Github Skills (14)
ccl10
c1710
cuda10
testing10
mpi10
c-language10
hpc10
openmpi10
parallel-computing10
c1110
cprogramming-language10
performance-testing10
performance-optimization9
fortran9
Programming languages (4)
Github contributions (5)
NVIDIA/nccl-tests
Aug 2017 - Jan 2023
NCCL Tests
Role in this project:
QA Engineer / Test Automation Engineer
QA Engineer / Test Automation EngineerContributions:8 reviews, 27 commits, 9 PRs in 5 years 5 months
Contributions summary:Steve primarily contributed to the testing framework within the NCCL tests repository. Their commits added a new alltoall performance test, including the necessary initialization and execution logic. They also addressed a compilation issue related to NCCL version compatibility and modified bandwidth calculation for specific collectives. Furthermore, they updated the NCCL tests suite, introducing features like timeout handling, improved CPU time reporting, and support for CUDA graph launches.
gpucudampinccl
NVIDIA/nccl
Jan 2016 - Dec 2022
Optimized primitives for collective multi-GPU communication
Role in this project:Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:11 reviews, 112 commits, 90 PRs in 6 years 11 months
Contributions summary:Steve primarily focused on improving and debugging the MPI test suite, which tests the collective communication primitives. They moved test files, improved the output format, and fixed bugs. Additionally, the user addressed potential deadlocks in the core communication routines and removed unneeded includes, indicating work on optimizing the library's performance. The user also made improvements to the low level device level operations and device/host memory interactions.
cudampiinfinibandgpucluster-computing
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.