Milad Mohammadi
Software Engineer Lead PyTorch at Meta
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMilad Mohammadi is a senior software engineering leader and ML systems expert with 11+ years of experience building high-performance AI runtimes, compilers, and distributed compute platforms from research to production. Currently leading PyTorch at Meta after scaling PyTorch TPU efforts at Google, he has shipped TPU support across multiple generations and driven features like quantization, FP8, gSPMD and custom kernels. His background spans Stanford PhD work in energy-efficient compilers and architecture, leadership roles at Apple and startups, and founding a clinical AI company—bringing both deep technical rigor and product/entrepreneurial instincts. An active open-source contributor, he has added core operations and shape-inference improvements to PyTorch/XLA, helping enable PyTorch on TPUs at cloud scale. Notably, he blends compiler-level systems know-how with hands-on ML engineering, making him adept at moving research-grade innovations into robust, large-scale tooling.
11 years of coding experience
6 years of employment as a software developer
Certificate Mini-MBA Innovation & Entrepreneurship, Certificate Mini-MBA Innovation & Entrepreneurship at Stanford University Graduate School of BusinessBS (Honors) Electrical Engineering, BS (Honors) Electrical Engineering at The University of British ColumbiaPh.D. Computer Systems, Ph.D. Computer Systems at Stanford University
Github Skills (17)
pytorch10
c-language10
operation10
python10
tensorrt10
deep-learning10
tensorflow10
xla10
tensor10
cprogramming-language10
linear-algebra10
machine-learning9
autograd9
compiler-compiler9
compiler9
Programming languages (6)
Github contributions (5)
pytorch/xla
Jun 2021 - Dec 2022
Enabling PyTorch on XLA Devices (e.g. Google TPU)
Role in this project:
Back-end Developer
Back-end DeveloperContributions:3 releases, 422 reviews, 331 commits in 1 year 5 months
Contributions summary:Milad primarily focused on implementing and testing new PyTorch operations within the XLA framework. The commits demonstrate the addition of `std_mean`, `var_mean`, `amin`, `amax`, `sgn`, `sign`, `inverse`, and `logdet` operations, alongside associated testing code. These contributions involve modifying core XLA code, the tensor methods, and test files to integrate and validate these new functionalities, improving PyTorch/XLA's coverage.
pytorchxladeep-learningtpucompiler
pytorch/pytorch
May 2022 - Aug 2022
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:Backend Developer
Backend DeveloperContributions:43 reviews, 14 commits, 10 PRs in 3 months
Contributions summary:Milad's primary contributions focused on enhancing the shape inference capabilities within the LazyTensor framework. They implemented support for the `expand` operation, addressing related issues and integrating symbolic constants. Additionally, the user worked on updating the build process for PyTorch/XLA CI testing, and added methods to the `dynamic_ir` module. These changes involved modifications to both C++ and Python code, crucial for the overall functionality of the PyTorch library, specifically the XLA integration.
pythongpu-accelerationdeep-learninggpunumpy
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.