Evan Racah
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolEvan Racah is a machine learning software engineer with a decade of experience building production-ready ML systems and research-grade models across autonomous vehicles, large-scale training frameworks, and scientific computing. He holds a Master’s in Computer Science from Université de Montréal/Mila and has applied deep learning, computer vision, and reinforcement learning at Waymo Research and MosaicML, where he contributed concrete improvements such as implementing Adaptive Gradient Clipping in the popular MosaicML Composer. Comfortable bridging research and engineering, Evan has a history at national labs and startups—translating experiments into robust code and unit-tested features for real-world workloads. Based in San Francisco, he combines academic rigor from Mila with hands-on systems experience at Databricks and Berkeley Lab, and often focuses on making advanced ML algorithms reliably usable in production.
10 years of coding experience
9 years of employment as a software developer
Bachelor's Degree, Mechanical Engineering with a Minor in Computer Science, Bachelor's Degree, Mechanical Engineering with a Minor in Computer Science at University of California, DavisMaster's Degree, Computer Science, Master's Degree, Computer Science at Université de Montréal
English, Spanish, French
Stackoverflow
46reputation
194reached
1answer
0questions
Github Skills (8)
neural-network10
algorithms10
pytorch10
machine-learning10
deep-learning10
python10
testing10
jupyter-notebook6
Programming languages (5)
Github contributions (5)
mosaicml/composer
Apr 2022 - Jan 2023
Supercharge Your Model Training
Role in this project:
ML Engineer
ML EngineerContributions:758 reviews, 56 commits, 232 PRs in 8 months
Contributions summary:Evan implemented Adaptive Gradient Clipping (AGC) to the composer/algorithms module, developing both functional and trainer interfaces for its application. Their work included creating a functional interface (`apply_agc`) and a trainer interface (`AGC`), along with testing through example models like simple_model_with_grads and cnn_model_with_grads. The user also contributed to unit tests for various AGC scenarios, ensuring the accuracy and functionality of the implemented algorithm, specifically tested against 1D, 2D, 3D and 4D weights and gradients.
pytorchml-systemsdeep-learningneural-networksmachine-learning
eracah/slot-contrastive-networks
Sep 2019 - Jul 2020
Contributions:226 commits, 1 push, 1 branch in 10 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.