Ilia Cherniavskii

San Francisco, California, United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Ilia Cherniavskii is a machine learning engineer with eight years of experience, currently working on PyTorch at Facebook in San Francisco. He brings deep backend expertise in ML frameworks, having contributed to high-impact open-source projects like Kineto (CPU/GPU profiling) and the core Caffe2 framework, where he implemented backpropagation support for control-flow operators. His work demonstrates a practical understanding of low-level initialization, performance tracing, and build-time robustness—fixing segfault guards, clock discrepancies, and invalid JSON outputs to improve stability. Trained at Saint-Petersburg Academic University, he combines systems-level engineering with ML insight to make complex deep-learning infrastructure reliable and production-ready.

8 years of coding experience

Saint-Petersburg Academic University

English, Russian

Github Skills (16)

cuda10

pytorch10

c-language10

cprogramming-language10

backpropagation10

profiling10

python9

computer-engineering9

ai9

deeplearning-ai9

machine-learning9

deep-learning9

json9

build-system8

performance-optimization7

Programming languages (5)

C++ShellHTMLJupyter NotebookPython

Github contributions (5)

pytorch/kineto

Nov 2020 - Jul 2021

A CPU+GPU Profiling library that provides access to timeline traces and hardware performance counters.

Role in this project:

Back-end Developer

Contributions:1 review, 28 commits, 17 PRs in 7 months

Contributions summary:Ilia primarily contributed to the `kineto` library, focusing on integrating it with PyTorch. Their work included implementing integration features and adding guards to prevent startup seg faults, indicating an understanding of build processes. Furthermore, they exposed parameters and fixed issues, such as invalid JSON outputs and clock discrepancies, enhancing stability and functionality. The user's contributions touched upon core library components, specifically addressing initialization, activity profiling, and output mechanisms.

cudacputracesprofilingtimeline

facebookarchive/caffe2

Aug 2017 - Mar 2018

Caffe2 is a lightweight, modular, and scalable deep learning framework.

Role in this project:

Back-end Developer & ML Engineer

Contributions:69 commits, 4 PRs, 1 push in 6 months

Contributions summary:Ilia primarily contributed to the core deep learning framework Caffe2. Their work involved modifying core components, such as the workspace and argument definitions, to enhance functionality, especially related to control flow operators like "If" and "While". Key contributions include implementing backpropagation support for these operators, which involved changes to the Do operator and the generation of gradient operators. They also focused on improving the efficiency and structure of the framework by refactoring the executor and enhancing stream selection.

pytorchscalablecaffe2deep-learningml

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.

Request Free Trial