Artem Rakhov
United States
Summary
🤩
Rockstar🎓
Top SchoolArtem Rakhov is a seasoned software engineer with 8 years focused on ML frameworks and backend systems, with deep experience in autonomous vehicle stacks at Woven by Toyota and Lyft. He helped build Glow at Facebook/PyTorch, contributing operator support and graph optimizations that enabled models like VGG16 to run efficiently on hardware accelerators. Earlier work at Google centered on an IPython-based ML platform emphasizing model understandability, reflecting a consistent interest in making complex ML systems interpretable and production-ready. Artem combines low-level ML compiler fixes with large-scale autonomous systems engineering, bridging research-quality models and real-world deployment. Based in the United States, he brings a pragmatic engineering approach informed by both open-source contributions and production AV experience. An understated strength is his knack for subtle correctness fixes (e.g., image-loading and operator bugs) that unlock bigger performance and usability gains.
8 years of coding experience
11 years of employment as a software developer
Specialist Computer Science, Specialist Computer Science at Saratov State University named after N.G.Chernyshevsky
Github Skills (9)
convolution10
c-language10
cprogramming-language10
optimization9
optimisation9
graph9
deep-learning8
machine-learning8
onnx4
Programming languages (2)
Github contributions (3)
pytorch/glow
Dec 2017 - Aug 2019
Compiler for Neural Network hardware accelerators
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:177 commits, 142 PRs, 59 pushes in 1 year 8 months
Contributions summary:Artem implemented support for the LRN operator in the Caffe2 importer, enabling the model to be used in the Glow compiler. They also fixed issues in existing convolution and LRN operator implementations, allowing for the successful execution of the VGG16 model. Furthermore, the user updated the image loading function to prevent image rotation, and replaced the Slice of Splat node with a new Splat node during graph optimization, potentially improving performance.
hardware-acceleratorscompilerneural-networkacceleratorshardware
artemrakhov/glow
May 2018 - Aug 2019
Compiler for Neural Network hardware accelerators
Contributions:87 pushes, 88 branches in 1 year 3 months
hardware-acceleratorscompilerneural-networkacceleratorshardware
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.