Szymon Tworkowski
Member Of Technical Staff at xAI
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSzymon Tworkowski is a machine learning researcher and software engineer with seven years of experience focused on scaling and improving large language models. Currently a Member of Technical Staff at xAI and a PhD student at the University of Warsaw, he builds practical methods to extend LLM context and reasoning—most notably the Focused Transformer (LongLLaMA) project. His background includes research stints at Google and DeepMind, production internships at Amazon and Goldman Sachs, and significant open-source contributions to the Trax deep learning library (including Funnel Transformer work). Based in San Francisco, he combines deep academic training with hands-on engineering across model architecture, efficient implementation, and testing. Szymon’s long-term ambition is to create increasingly autonomous LLMs that can assist with and ultimately generate novel scientific knowledge.
7 years of coding experience
3 years of employment as a software developer
Doctor of Philosophy - PhD, Machine Learning, Doctor of Philosophy - PhD, Machine Learning at University of Warsaw
Polish, English, German, French, Chinese
Github Skills (11)
transformer10
machine-learning10
deep-learning10
jax10
python10
numpy9
tensorflow8
data-structure7
data-structures7
algorithm7
algorithms7
Programming languages (6)
Github contributions (5)
google/trax
Oct 2020 - Oct 2021
Trax — Deep Learning with Clear Code and Speed
Role in this project:
ML Engineer
ML EngineerContributions:8 reviews, 32 commits, 15 PRs in 11 months
Contributions summary:Szymon significantly contributed to the development of the Trax deep learning library, specifically focusing on the Funnel Transformer model. Their work involved implementing a customizable pooling layer, integrating an alternative funnel block, and adding unit tests. Further contributions included refactoring and restructuring the code, splitting and modifying pooling layers, and refining the model's internal components and tests, demonstrating a strong understanding of model architecture and implementation.
pytorchspeeddeep-learningnumpyreinforcement-learning
Vatican-X-Formers/trax
Oct 2020 - Jan 2022
Trax — Deep Learning with Clear Code and Speed
Contributions:17 reviews, 46 PRs, 351 pushes in 1 year 2 months
pytorchspeeddeep-learningmachine-learningclear
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.