Volodymyr Kysenko
Staff Software Engineer at Google
Sunnyvale, California, United States
Summary
🤩
Rockstar🎓
Top SchoolVolodymyr Kysenko is a Staff Software Engineer in Google’s On-Device Machine Learning group with 14 years of experience building high-performance image and ML inference systems. He has deep expertise in Halide and compiler backends, having led the Halide backend for Cadence DSPs and contributed core features like multidimensional vectorization and new language directives. His work on HDR+ camera pipelines and numerous ISP optimizations helped ship production-quality imaging on Pixel devices, and he continues to maintain Halide at Google. Comfortable across low-level performance engineering, OpenCL, and VLIW/embedded targets, he bridges research prototypes and production inference. An active contributor to the prominent halide/Halide repo, he focuses on targeted performance improvements and benchmarking for GPU/OpenCL targets. Based in Sunnyvale, he combines rigorous CS training with hands-on hardware-software co-design experience that often surfaces in subtle algorithmic and microarchitectural optimizations.
14 years of coding experience
9 years of employment as a software developer
Master's degree Computer Science, Master's degree Computer Science at Taras Shevchenko National University of Kyiv
English, Ukrainian
Github Skills (15)
performance-tuning10
opencl10
halide10
performance-analytics10
performance-monitor10
performance-analysis10
performance-monitoring10
performance-measurement10
benchmarking9
benchmark9
compiler-optimization9
c-language8
cprogramming-language8
gpu8
image-processing7
Programming languages (2)
Github contributions (5)
halide/Halide
May 2014 - Jan 2023
a language for fast, portable data-parallel computation
Role in this project:
Performance Engineer
Performance EngineerContributions:268 reviews, 630 commits, 179 PRs in 8 years 10 months
Contributions summary:Volodymyr's commits primarily focus on enhancing the performance of the Halide library through targeted optimizations and the addition of performance tests. Their contributions involve implementing new functions for specific OpenCL targets to improve computational efficiency, such as adding minval\_f32/maxval\_f32 functions. They also addressed style corrections and fixed a typo related to GPU operations, while also contributing to the integration of matrix multiplication and merge sort for benchmarking.
computationhexagonhalideparallelgpu
vksnk/go-fann
Jan 2013 - Feb 2015
Contributions:5 commits, 1 push, 2 comments in 2 years 1 month
golangartificialneural-networksneural-networkartificial-neural-networks
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.