Szymon Migacz
Distinguished Engineer at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolSzymon Migacz is a Distinguished Engineer at NVIDIA with 11 years of experience advancing deep learning and CUDA software, rising through roles from CUDA library developer to principal and now senior technical leader. He combines a strong theoretical background in physics and mathematics (MSc, University of Warsaw) with hands-on expertise optimizing reduced-precision training and production-grade deep learning models. At NVIDIA he has driven algorithmic improvements and performance work across libraries and model implementations, including contributions to the high-profile NVIDIA/DeepLearningExamples repository on GNMT and Transformer-XL where he improved AMP training and gradient accumulation. Based in California, he is known for bridging low-level CUDA numerics and higher-level ML training efficiency, enabling faster, more accurate model training on enterprise GPU infrastructure.
11 years of coding experience
11 years of employment as a software developer
Master of Science (MSc) Theoretical physics Mathematical physics, Master of Science (MSc) Theoretical physics Mathematical physics at University of Warsaw
English, Polish
Github Skills (6)
pytorch10
machine-translation10
deep-learning10
nlp9
computer-vision3
tensorflow3
Programming languages (3)
Github contributions (5)
NVIDIA/DeepLearningExamples
Aug 2018 - Jun 2022
State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and performance on enterprise-grade infrastructure.
Role in this project:
ML Engineer
ML EngineerContributions:18 commits, 51 PRs, 42 pushes in 3 years 11 months
Contributions summary:Szymon primarily contributes to the PyTorch implementation of the GNMT model for machine translation. Their work involves adding and updating the GNMT model, including modifications to the training script and configurations. They also addressed issues and made updates to the Transformer-XL model, specifically regarding AMP training and gradient accumulation, suggesting a focus on improving the efficiency and performance of deep learning models within the repository.
forecastingcaffe2translationspeech-recognitionstate-of-the-art
szmigacz/qmk_firmware
Jun 2019 - Oct 2024
Contributions:6 pushes, 2 branches in 5 years 5 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.