David Mizrahi
Member Of Technical Staff at Anthropic
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolDavid Mizrahi is a Machine Learning engineer with six years of experience building and scaling vision and ML systems, currently on the technical staff at Anthropic after research roles at Apple. He holds a Master’s in Data Science and a Bachelor's in Computer Science from EPFL, combining strong academic foundations with hands-on engineering. David has contributed to prominent open-source computer vision work—integrating Swin Transformer (Swin-B, Swin-L) and DDP support into the well-known OpenPifPaf project—demonstrating expertise in model architecture, distributed training, and production-ready refactors. His background blends research and product-focused development, moving transformer-based CV advances from checkpoints to usable toolchains. Based in San Francisco, he thrives at the intersection of cutting-edge research and deployable ML infrastructure.
6 years of coding experience
2 years of employment as a software developer
Master's degree, Data Science, Master's degree, Data Science at EPFL
Github Skills (10)
swin-transformer10
transformer-models10
computer-vision10
pytorch10
deep-learning10
python10
pose-estimation9
checkpointing9
model-checking9
checkpoint9
Programming languages (4)
Github contributions (5)
openpifpaf/openpifpaf
Aug 2021 - Sep 2021
Official implementation of "OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association" in PyTorch.
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 5 commits, 5 PRs in 1 month
Contributions summary:David made significant contributions to integrating and enhancing Swin Transformer, a transformer-based architecture for computer vision, into the OpenPIFPAF project. This involved implementing new Swin Transformer architectures (Swin-B, Swin-L) and modifying the existing code to accommodate larger models and DDP (Distributed Data Parallelism) compatibility. The user also focused on refactoring, CLI argument changes, and adding pretrained checkpoints for the newly integrated Swin Transformer models.
pytorchkeypoint-detectionopenpifpafassociationdeep-learning
EPFL-VILAB/MultiMAE
Mar 2022 - Dec 2022
Contributions:11 commits, 9 pushes, 10 comments in 8 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.