Laith Sakka
Staff Research Scientist at Meta
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolLaith Sakka is a Staff Research Scientist at Meta with a decade of experience building high-performance systems, compilers, and ML infrastructure. He earned a PhD from Purdue for compiler optimizations and has published PLDI and OOPSLA work on program transformation and locality, then drove major Velox and PyTorch efforts at Meta to boost execution speed, reliability, and developer ergonomics. Laith’s contributions to PyTorch’s Dynamo and to Velox’s function authoring and fuzzing frameworks show a rare blend of deep compiler expertise and practical ML-engine engineering. He excels at turning research into production—whether fusing recursive traversals for 5× speedups or hardening large-scale engines with automated validation—and is active in the PyTorch open-source community. Based in Seattle, he pairs rigorous academic training with a track record of shipping performance-critical infrastructure at scale.
10 years of coding experience
7 years of employment as a software developer
Bachelor's degree Computer Engineering, Bachelor's degree Computer Engineering at Princess Sumaya University for TechnologyDoctor of Philosophy (PhD) Computer Engineering, Doctor of Philosophy (PhD) Computer Engineering at Purdue University
English, Arabic
Github Skills (12)
amazon-dynamodb10
pytorch10
machine-learning10
deep-learning10
dynamodb10
python10
aws-dynamodb10
autograd10
testing9
neural-network9
tensor8
gpu7
Programming languages (7)
Github contributions (5)
pytorch/pytorch
Jan 2024 - Apr 2025
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
ML Engineer
ML EngineerContributions:352 reviews, 226 PRs, 1824 pushes in 1 year 2 months
Contributions summary:Laith primarily contributed to the PyTorch library, focusing on the Dynamo component. Their work involved implementing and fixing functionalities related to automatic differentiation and optimization. This includes addressing issues with operator support, improving the handling of random calls within inlined code, and extending support for built-in callables. Additionally, the user contributed tests to ensure the correctness of these features.
pythongpu-accelerationdeep-learninggpunumpy
laithsakka/pytorch
Jan 2024 - Jan 2025
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:174 pushes, 34 branches in 1 year
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.