Grigory Sizov
Staff Software Engineer, Machine Learning at Meta
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolGrigory Sizov is a Staff Software Engineer in Machine Learning at Meta with nine years of experience optimizing large language model inference and other AI workloads on GPUs. He contributes to AITemplate and the PyTorch ecosystem (Core, TorchAudio, TorchRL), notably integrating WavLM into pytorch/audio with TorchScript and quantization tests and performance-focused refactors. His background spans industry and research—from delivering ML products and platforms at Bain to a PhD and postdoc in theoretical physics—giving him deep mathematical rigor paired with production engineering chops. Based in London, he combines low-level GPU performance tuning with high-level model integration, and is an active open-source maintainer and advocate for reproducible ML tooling. An unexpected thread through his career is applied physics research informing robust, numerically stable ML implementations.
9 years of coding experience
7 years of employment as a software developer
Master of Science (M.S.) Applied Mathematics and Physics, Master of Science (M.S.) Applied Mathematics and Physics at Moscow Institute of Physics and Technology (State University) (MIPT)Doctor of Philosophy (PhD) Theoretical and Mathematical Physics, Doctor of Philosophy (PhD) Theoretical and Mathematical Physics at King's College LondonMaster of Science (M.S.) Physics, Master of Science (M.S.) Physics at University of Waterloo
English, Russian, French, Portuguese, German
Stackoverflow
99reputation
135kreached
2answers
0questions
Github Skills (14)
machine-learning10
pytorch10
audio-processing10
python10
huggingface-transformers9
torchscript8
audio7
quantization7
pandas6
date-time-conversion6
tensorboard6
tensorflow6
numpy6
timestamp6
Programming languages (5)
Github contributions (5)
pytorch/audio
Nov 2022 - Jan 2023
Data manipulation and transformation for audio signal processing, powered by PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:24 reviews, 6 commits, 6 PRs in 1 month
Contributions summary:Grigory contributed significantly to the `pytorch/audio` repository by implementing and integrating the WavLM model. They added the `WavLMSelfAttention` class, factory functions, and WavLM model bundles, which involved adjusting existing components to be compatible. Furthermore, the user expanded HuggingFace integration tests and implemented TorchScript and quantization tests for the WavLM model. The user also refactored code to utilize `torch.nn.MultiheadAttention` (BetterTransformer) for improved performance and added HiFi GAN generator to prototypes.
pytorchaudio-processingmanipulationpythonsignal
sgrigory/audio
Nov 2022 - Jun 2023
Data manipulation and transformation for audio signal processing, powered by PyTorch
Contributions:72 pushes, 4 branches in 7 months
pytorchmanipulationsignaltransformationmachine-learning
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.