Berton Earnshaw
Research Professor, Department Of Mathematics And School Of Biological Sciences
Salt Lake City, Utah, United States
Summary
👤
Senior🎓
Top SchoolBerton Earnshaw is a research professor and machine learning practitioner with 12 years of experience applying mathematical rigor to biological problems from Salt Lake City. He blends academic leadership at the University of Utah with industry-facing roles at Recursion and Valence Labs, advancing ML-driven discovery in life sciences. His hands-on contributions include integrating the RxRx1 high-content imaging dataset into the widely used WILDS benchmark, demonstrating expertise in dataset engineering, preprocessing, and evaluation for real-world distribution shifts. Trained with a PhD in mathematics, he moves comfortably between theory and production, optimizing pipelines and learning schedules to improve model robustness. Colleagues value his rare combination of deep mathematical background, practical ML engineering, and sustained focus on translational impact.
12 years of coding experience
17 years of employment as a software developer
The University of UtahMS Mathematics, MS Mathematics at Brigham Young University
Portuguese, Spanish
Github Skills (12)
computer-vision10
data-loading10
eval10
machine-learning10
pytorch10
python10
evaluation10
image-processing10
datasets10
data-set10
resnet9
transformers8
Programming languages (7)
Github contributions (5)
p-lambda/wilds
Mar 2021 - May 2021
A machine learning benchmark of in-the-wild distribution shifts, with data loaders, evaluators, and default models.
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 9 commits, 1 PR in 1 month
Contributions summary:Berton significantly contributed to integrating the RxRx1 dataset into the `wilds` benchmark. This involved creating a custom dataset class (`RxRx1Dataset`), defining data splits, and implementing the necessary data loading and transformation logic, including image preprocessing and standardization. Additionally, the user configured the dataset within the example configurations, set up relevant transforms, and specified evaluation metrics for the dataset. Furthermore, the user updated the download URL, added validation set and metadata, and adjusted the learning rate scheduler for the RxRx1 dataset.
loadersshiftswildmachine-learningmachine-learning-benchmark
bearnshaw/dotvim
Nov 2014 - Apr 2020
Contributions:44 commits, 27 pushes, 2 branches in 5 years 6 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.