David Olsen
Software Engineer at NVIDIA / PGI
Portland, Oregon, United States
Summary
🤩
RockstarDavid Olsen is a software engineer with 8 years of experience specializing in compiler back ends and CUDA-enabled high-performance C++ libraries. Based in Portland, he contributes to prominent open-source projects at NVIDIA and LLVM, improving CUDA backends, ClangIR code generation, and cross-compiler compatibility. His work shows deep familiarity with compiler internals, PTX/kernel dispatch, and platform-agnostic build configurations, often resolving subtle regressions and floating-point issues that break builds. David combines pragmatic maintenance—fixing warnings and preventing compilation errors—with feature work like integral and boolean support in ClangIR. He brings a practitioner’s eye for production robustness in HPC toolchains, and his contributions to well-known repos such as thrust, cub, and llvm reflect both depth and impact.
8 years of coding experience
Github Skills (19)
thrust10
algorithms10
c-language10
gpgpu10
llvm10
mlr10
nvidia10
compiler-design10
cub10
gpu10
cubit10
accelerated-computing10
cuda10
cpp10
cprogramming-language10
Programming languages (8)
Github contributions (5)
NVIDIA/cub
Mar 2020 - Oct 2021
[ARCHIVED] Cooperative primitives for CUDA C++. See https://github.com/NVIDIA/cccl
Role in this project:
Back-end Developer
Back-end DeveloperContributions:6 reviews, 10 commits, 3 PRs in 1 year 7 months
Contributions summary:David primarily contributed to the CUDA C++ library `cub`, focusing on improving its functionality and compatibility. Their work involved modifying device code, implementing kernel dispatch configurations for different PTX versions, and addressing potential issues related to compilation with different compilers like NVC++. The user's commits show a deep understanding of the CUDA programming model, including kernel launches and device code execution. Furthermore, they were involved in optimizing code and resolving issues related to floating-point representation.
cxxprimitivesnvidia-hpc-sdknvidiacpp20
NVIDIA/thrust
Aug 2019 - Jun 2022
[ARCHIVED] The C++ parallel algorithms library. See https://github.com/NVIDIA/cccl
Role in this project:Back-end Developer
Back-end DeveloperContributions:8 reviews, 16 commits, 9 PRs in 2 years 10 months
Contributions summary:David primarily focused on modifying and maintaining the CUDA back end for the thrust library, a C++ parallel algorithms library. Their contributions involved addressing compiler warnings, fixing regressions, and improving code compatibility with different compilers and build configurations (including nvc++ and -stdpar). They also worked on ensuring the library's functionality and preventing compilation errors. Their changes indicate a focus on ensuring the library's correct behavior and platform-agnostic execution.
c-plus-pluscxxnvidia-hpc-sdknvidiacpp20
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.