Vitaly Fedyunin
Software Engineer at Waymo
New Providence, New Jersey, United States
Summary
🤩
Rockstar🎓
Top SchoolVitaly Fedyunin is a software engineer with 11 years of experience building high-performance data and infrastructure systems, currently focused on performance optimization at Waymo. He spent four years at Meta working on PyTorch optimization and data loading, and is an active contributor to the core PyTorch data repo where he improved DataLoader2 with checkpointing, prefetching, and multi-format file loaders to make large-scale ML data pipelines more efficient and robust. His background includes engineering roles at Google and senior technical leadership at ad-tech firms, giving him a rare mix of deep systems expertise and product-driven execution. Based in New Providence, NJ, he combines hands-on backend development with a strong track record of shipping measurable performance improvements in production ML stacks.
11 years of coding experience
11 years of employment as a software developer
Computer Science, Computer Science at Tomsk State University
Stackoverflow
26reputation
688reached
2answers
1question
Github Skills (21)
pytorch10
data-pipelines10
python10
data-loading10
data-pipeline10
caching9
multithreading9
hyper-threading9
thread9
threaded9
multiprocessing9
file-processing8
file-handling8
file-access8
fileio8
Programming languages (6)
Github contributions (5)
pytorch/data
Feb 2021 - Oct 2022
A PyTorch repo for data loading and utilities to be shared by the PyTorch domain libraries.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:118 reviews, 96 commits, 32 PRs in 1 year 8 months
Contributions summary:Vitaly primarily contributed to the core data loading and utility features of the PyTorch data repository. Their work involved significant modifications to the `DataLoader2` class, including incorporating features for checkpointing and reading services. The commits demonstrate an understanding of data pipeline design, with the implementation of new functionalities such as prefetching and integrating various file loaders for handling diverse data formats. The user also focused on improving data processing efficiency and correctness, including the introduction of new functionalities within the data pipeline.
pytorchpythondeep-learningtorchmachine-learning
VitalyFedyunin/pytorch
Dec 2018 - Jun 2022
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:348 pushes, 110 branches, 135 issues in 3 years 6 months
pythongpu-accelerationdeep-learninggpuacceleration
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.