Ilya Sherstyuk
Software Engineer - Deep Learning Inference Workflows at NVIDIA
Hawaii, United States
Summary
🤩
Rockstar🎓
Top SchoolIlya Sherstyuk is a software engineer specializing in deep learning inference workflows, currently contributing to NVIDIA’s TensorRT efforts to optimize model deployment and runtime efficiency. With seven years of experience and a BS in Computer Science from Caltech, he bridges research and production—having interned in quantitative engineering at Citadel and built embedded firmware and memory-efficient services at Verkada. He is an active ML systems contributor to the flagship PyTorch project, improving ONNX export accuracy and memory usage for dynamic operators like aten::slice and Softmax constant folding. Based in Hawaii, Ilya combines low-level systems thinking with practical tooling for ML inference, and his background in astronomical signal processing and teaching hints at a methodical, data-driven approach to problem solving.
7 years of coding experience
2 years of employment as a software developer
High School Diploma, High School Diploma at Punahou SchoolCalifornia Institute of Technology
Russian, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (8)
pytorch10
machine-learning10
onnx10
python10
deep-learning9
tensor8
autograd7
gpu6
Programming languages (3)
Github contributions (5)
pytorch/pytorch
Feb 2023 - Aug 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
ML Engineer
ML EngineerContributions:3 reviews, 10 PRs, 33 comments in 6 months
Contributions summary:Ilya primarily contributed to the ONNX exporter for PyTorch. Their work focused on refining and optimizing the export process, specifically addressing issues related to dynamic axes in the `aten::slice` operation, constant folding for the Softmax operation, and improving overall memory usage during export. They also implemented and tested new functionalities to enhance the accuracy and efficiency of ONNX models.
pythongpu-accelerationdeep-learninggpunumpy
ME-134/PuzzleBot
Jan 2022 - Mar 2022
Contributions:2 releases, 131 commits, 98 pushes in 2 months
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.