Eugene Chereshnev
Software Engineer at Intel Corporation
Hillsboro, Oregon, United States
Summary
🤩
Rockstar🎓
Top SchoolEugene Chereshnev is a software engineer with 11 years of experience at Intel, based in Hillsboro, Oregon, combining deep numerical and systems expertise with production-grade reliability. He holds a BS and MS in Computer Science from Novosibirsk State University and focuses on backend and performance engineering work for high-performance libraries. Eugene has contributed to flagship open-source projects like LAPACK and oneDNN, fixing intricate Fortran/math bugs and hardening benchmarking code to prevent out-of-bounds memory issues and improve reproducible performance results. His strengths lie in debugging complex numerical routines, optimizing memory and benchmark stability, and ensuring correctness in performance-critical code. Colleagues rely on him to find subtle mathematical and memory-access errors that elude typical testing, a skill reflected in his contributions to widely used scientific libraries. He blends academic grounding with practical engineering to keep demanding computational stacks both fast and correct.
11 years of coding experience
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at Novosibirsk State University (NSU)
Github Skills (13)
fortran10
memory-management10
matrix-factorization10
benchmarking10
c-language10
benchmark10
cprogramming-language10
performance-optimization10
lapack10
linear-algebra10
blas9
opencl9
eigenvalue8
Programming languages (3)
Github contributions (5)
uxlfoundation/oneDNN
Dec 2018 - Jan 2023
oneAPI Deep Neural Network Library (oneDNN)
Role in this project:
Backend & Performance Engineer
Backend & Performance EngineerContributions:189 reviews, 1363 commits, 66 PRs in 4 years 2 months
Contributions summary:Eugene primarily contributed to the performance and correctness of the library. Their commits show a focus on optimizing and refining existing functions within the "benchdnn" test suite. The user was involved in adding memory protection, fixing access issues, outputting engine types, and addressing other potential out-of-bounds memory accesses, improving stability and accuracy of the profiling. These changes included modifications to benchmark settings to ensure accurate, reproducible performance results.
bfloat16sse41avx512openmpopencl
Reference-LAPACK/lapack
Aug 2016 - Dec 2022
LAPACK development repository
Role in this project:Backend Developer
Backend DeveloperContributions:43 commits, 11 PRs, 5 comments in 6 years 4 months
Contributions summary:Eugene primarily contributed to the LAPACK development repository by fixing incorrect function calls and addressing issues within the Fortran source code. Their work involved identifying and correcting errors in mathematical functions, specifically related to complex matrix operations like GEMM and those used in singular value decomposition (SVD). The user also addressed issues in existing code, improving the functionality and accuracy of the LAPACK library.
eigenvalueslapacklapackelinear-algebrasingular-values
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.