Jamie Simon
Research Fellow at University of California, Berkeley
Reston, Virginia, United States
Summary
👤
Senior🎓
Top SchoolJamie Simon is a research-focused machine learning scientist and PhD candidate in Physics at UC Berkeley with a decade of experience probing the first-principles dynamics and generalization of deep neural networks. As a Research Fellow at Imbue (after an internship there), Jamie has contributed to foundational theory of training dynamics for self-supervised learning and studies neural net dynamics with an eye toward principled understanding rather than empirical tuning. He is also an active open-source contributor—helping expand BIG-bench with logical reasoning and sequence-manipulation tasks and building language-specific augmentations for NL-Augmenter—showing a blend of theory and practical tooling. Based in Reston, VA, Jamie combines deep physics training and hands-on ML engineering to tackle interpretable, theory-driven problems in modern AI, often translating theoretical insights into reproducible benchmarks and filters used by the community.
10 years of coding experience
2 years of employment as a software developer
Doctor of Philosophy - PhD, Physics, Doctor of Philosophy - PhD, Physics at University of California, BerkeleyVirginia Tech
Github Skills (10)
workbench10
machine-learning10
nlp10
testbench10
python10
text-classification9
regular-expression9
json9
open-source8
numpy8
Programming languages (4)
Github contributions (5)
google/BIG-bench
Mar 2021 - Jul 2021
Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models
Role in this project:
ML Engineer
ML EngineerContributions:52 commits, 5 PRs, 56 comments in 4 months
Contributions summary:Jamie contributed to the `big-bench` repository, which focuses on language model benchmarking. Their changes involved modifying the `LogicalDeductionTask` and the creation of the `sequence_manipulation` task. These changes suggest a focus on expanding the benchmark tasks, likely involving the development or refinement of logical reasoning and sequence manipulation assessments for language models. These code modifications indicate a direct engagement with the core functionality of the benchmark.
bertmachine-learningbenchmarkmeasuringbenchmarks
GEM-benchmark/NL-Augmenter
Sep 2021 - Sep 2021
NL-Augmenter 🦎 → 🐍 A Collaborative Repository of Natural Language Transformations
Role in this project:Back-end Developer
Back-end DeveloperContributions:7 commits, 1 PR, 9 comments in 15 days
Contributions summary:Jamie primarily focused on developing a filtering mechanism for text within the NL-Augmenter repository. They implemented a "Britishness" filter, which involved defining British English vocabulary and slang, and integrated it into the project's architecture. The user further refined this filter, renaming it to "Englishness" and addressing minor typos within the code. Their work contributes to the project's ability to augment text with language-specific variations.
nlpinformation-theorysentencedeep-learningmachine-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.