Vladimir Cherepanov
Software Engineer at NVIDIA
San Jose, California, United States
Summary
👤
Senior🎓
Top SchoolVladimir Cherepanov is a software engineer with two decades of broad technical experience and a current focus on deep learning frameworks at NVIDIA. He has a strong background in high-performance C++ and backend systems from roles at VMware, Yandex, and earlier enterprise and embedded software positions, and brings nine years of recent hands-on industry experience. His open-source contributions to Apache MXNet include optimizing CachedOp graph transformations and adding AMP support for NumPy ops—work that improves performance and numeric handling in a widely used deep learning framework. Based in San Jose with an MS in Computer Science from Bauman Moscow State Technical University, he blends low-level systems expertise with ML framework internals. Known for untangling tricky backward-pass and merge issues, he excels at making complex graph and performance problems practical for production.
9 years of coding experience
20 years of employment as a software developer
MS in Computer Science, computer simulation, MS in Computer Science, computer simulation at Bauman Moscow State Technical University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (12)
mxnet10
deep-q-learning10
c-language10
deep-learning10
cprogramming-language10
cudnn10
cuda9
numerics8
linear-algebra8
numeric8
numerical8
computation8
Programming languages (3)
Github contributions (5)
apache/mxnet
Jun 2020 - Dec 2021
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:
Back-end Developer
Back-end DeveloperContributions:8 reviews, 11 commits, 12 PRs in 1 year 6 months
Contributions summary:Vladimir primarily focused on optimizing the `CachedOp` in the MXNet framework, a core component for graph transformations. They made changes to allow input reordering during Gluon graph transformations and fixed issues related to backward passes and merge errors within the `CachedOp`. Furthermore, the user contributed to AMP (Automatic Mixed Precision) support within the NumPy operations, adding support and addressing handling issues. They also made further improvements related to CUdnn batch norm layers.
pythonschedulerdataflowmutationdata-science
mk-61/incubator-mxnet
Apr 2020 - Dec 2021
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Contributions:103 pushes, 13 branches in 1 year 9 months
pythonschedulerdataflowmutationorchestration
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.