Ivan Kobzarev
Software Engineer at Meta
Munich, Bavaria, Germany
Summary
🤩
Rockstar🎓
Top SchoolIvan Kobzarev is a software engineer with 11+ years of experience specializing in high-performance ML infrastructure and mobile AI, currently contributing to PyTorch from Munich. He has driven performance and compiler-focused improvements in core PyTorch (including Inductor, JIT, and sharded tensor work) and helped bring ML models to Android through PyTorch Mobile and demo apps. Long experience in Android at Meta and Odnoklassniki combines with earlier server-side Java and QA work, giving him a strong cross-stack perspective on optimization and reliability. Notably, he contributed pragmatic enhancements to TorchRec executors and PlayTorch mobile ML pipelines, showing an eye for both systems-level performance and hands-on mobile integration.
11 years of coding experience
11 years of employment as a software developer
Master's degree Computer science parallel algorithms, Master's degree Computer science parallel algorithms at Saint Petersburg State University
Русский, Английский
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (34)
pytorch10
c-language10
recommendation-system10
mobile-development10
python10
operation10
image-processing10
tensorrt10
machine-learning10
audio-processing10
android-development10
java10
javas10
deeplearning-ai10
deep-learning10
Programming languages (9)
Github contributions (5)
pytorch/android-demo-app
Sep 2019 - Jul 2020
PyTorch android examples of usage in applications
Role in this project:
Mobile Developer (Android)
Mobile Developer (Android)Contributions:62 reviews, 26 commits, 34 PRs in 9 months
Contributions summary:Ivan primarily contributed to the development of an Android application, as evidenced by the file paths and code changes within the `pytorch/android-demo-app` repository. They implemented features like image classification using PyTorch models, including incorporating a quantized model and UI improvements. The user also worked on refactoring code for efficient input tensor handling and integrated a moving average calculation for performance metrics, demonstrating a focus on optimization.
pytorchandroidandroid-examples
pytorch/pytorch
Aug 2019 - Jan 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:178 reviews, 366 commits, 432 PRs in 3 years 6 months
Contributions summary:Ivan's commits primarily focus on enhancing the performance and functionality of the PyTorch library. Their work includes optimizing and fixing code related to the JIT compiler, specifically addressing issues around Awaitable types and improving the handling of subclasses within the Inductor framework. They also contribute to the sharded tensor functionality, and add performance-related improvements within symbolic shapes, and improve the handling of the with_effects calls within Inductor. The user also added support for the addition of several functional collectives to dynamo remapping.
pythongpu-accelerationdeep-learninggpunumpy
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.