Alexey Varyzgin
Software Engineer Sr. at Intel Labs
Nizhny Novgorod, Nizhny Novgorod Oblast, Russia
Summary
🤩
RockstarAlexey Varyzgin is a senior software engineer with seven years of experience, currently working at Intel Labs in Nizhny Novgorod. He specializes in high-performance CPU backend development for AI inference, contributing notable BF16 (bfloat16) optimizations and mixed-precision support to the widely used open-source OpenVINO toolkit. His work spans layer-level enhancements (Gemm/MatMul, Deconvolution, MVN, Interpolate), AVX512 BF16 simulation, and robustness improvements such as input validation. Comfortable in low-level performance engineering, he combines practical production focus with open-source collaboration at a project level that directly impacts AI deployment efficiency.
7 years of coding experience
Github Skills (8)
c-language10
deep-learning10
inference10
cprogramming-language10
optimisation10
openvino10
optimization10
memory-management9
Programming languages (1)
Github contributions (5)
openvinotoolkit/openvino
Jun 2020 - Mar 2022
OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference
Role in this project:
Backend Developer
Backend DeveloperContributions:88 reviews, 23 commits, 46 PRs in 1 year 9 months
Contributions summary:Alexey primarily focused on optimizing and expanding the OpenVINO toolkit's capabilities on the CPU backend, specifically targeting BF16 (bfloat16) support. Their contributions included enabling BF16 optimizations for various layers like Gemm/MatMul and Deconvolution, as well as incorporating BF16 support into the MVN (Mean Variance Normalization) and Interpolate layers. Furthermore, they addressed issues related to mixed-precision inference and introduced BF16 simulator for AVX512. They also made improvements to the existing CPU implementation and added input checks to the layers.
inference-enginepytorchmodel-optimizerdeep-learninggpu
alexey-varyzgin/openvino
May 2020 - Apr 2022
OpenVINO™ Toolkit - Deep Learning Deployment Toolkit repository
Contributions:2 reviews, 6 PRs, 267 pushes in 1 year 10 months
pytorchdeep-learningdeploymentopenvino-toolkitinference
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.