Stanislav Ivashkevich
Software Development Engineer 2 at Amazon Web Services (AWS)
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolStanislav Ivashkevich is a software engineer with 11 years of experience specializing in compiler backends and GPU performance, currently working as an SDE II at AWS in San Jose. He has deep hands‑on expertise from AMD and open-source LLVM/Clang contributions, where he implemented AMDGPU intrinsics, wave-barrier support, and byte selection optimizations. As a performance engineer on AMD’s MIOpen, he drove kernel-level improvements—replacing costly operations and restructuring loops—to deliver measurable convolution speedups. His academic background spans Waterloo, Toronto, and MSU, blending rigorous math and CS training with practical systems work. Colleagues would describe him as a pragmatic optimizer who finds elegant low-level fixes that unlock real-world performance gains.
11 years of coding experience
6 years of employment as a software developer
Lomonosov Moscow State UniversityBachelor’s Degree, Computer Science, Bachelor’s Degree, Computer Science at University of TorontoMaster of Mathematics, Computer Science, Master of Mathematics, Computer Science at University of Waterloo
Github Skills (29)
opencl10
c-language10
gpgpu10
llvm10
c1110
c1710
kernel10
compiler-design10
compiler-compiler10
gpu10
performance-optimization10
optimisation10
convolution10
compiler10
amdgpu10
Programming languages (7)
Github contributions (5)
llvm/llvm-project
Jan 2023 - Jan 2023
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:584 reviews, 8 commits, 221 PRs in 9 days
Contributions summary:Stanislav's contributions primarily involve modifying and enhancing the LLVM compiler infrastructure. Their work includes adding support for 6-element vectors in intrinsics, supporting byte_sel modifiers for AMDGPU instructions, and fixing a crash in the SILoadStoreOptimizer. They also made changes to add and improve functionalities related to byte selection for AMDGPU and implement new target intrinsics like `s_prefetch_data` and `s_buffer_prefetch_data`. The user's contributions also include improving the code generation of AMDGPU instructions, modifying and expanding the code base for better support and optimization.
compilerstechnologiesclangsubmittoolchain
ROCm/MIOpen
Mar 2017 - Apr 2017
AMD's Machine Intelligence Library
Role in this project:Performance Engineer
Performance EngineerContributions:25 commits in 1 month
Contributions summary:Stanislav focused on optimizing the performance of the MIOpen library's kernels, specifically for convolution operations. They identified and fixed performance regressions in existing kernels and improved overall efficiency. Their contributions include code restructuring, loop optimization, and replacing computationally expensive operations (like FP division) with more efficient alternatives (reciprocal + multiply), leading to significant performance gains for various backward and forward convolution scenarios.
amdcluster-computingmachine-learningblasmachine-intelligence
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.