Michael Lazos
Software Engineer at Meta
Greater Seattle Area United States
Summary
🤩
Rockstar🎓
Top SchoolMichael Lazos is a software engineer with 11 years of experience specializing in optimizing machine learning models for specialized hardware, FPGA microarchitecture, and compiler toolchains. Currently at Meta, he works on PyTorch compilers, Python bytecode simulation, and Triton/CUTLASS kernel generation and auto-tuning to bridge high-level ML workflows with high-performance kernels. Previously at Microsoft Research he prototyped FPGA implementations of ML models and built a compiler targeting an FPGA soft processor, bringing practical hardware-software co-design experience. He has contributed deep fixes and optimizer/torch-function work to the core PyTorch repository, demonstrating an uncommon grasp of both framework internals and low-level performance tuning. Based in the Greater Seattle Area with a Computer Engineering degree from Brown, Michael combines research rigor with production engineering to push ML models onto unconventional runtimes. He pairs a systems-oriented mindset with hands-on toolchain work—often focusing on the subtle details that make compiler-generated kernels actually fast on real devices.
11 years of coding experience
7 years of employment as a software developer
Bachelor of Science (BS) Computer Engineering, Bachelor of Science (BS) Computer Engineering at Brown University
English, Spanish, Greek
Github Skills (7)
automatic-differentiation10
pytorch10
machine-learning10
python10
c-language4
cuda4
cprogramming-language4
Programming languages (3)
Github contributions (5)
pytorch/pytorch
Aug 2022 - Jan 2023
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1221 reviews, 207 commits, 558 PRs in 5 months
Contributions summary:Michael made significant contributions to the PyTorch framework, particularly in the area of automatic differentiation and optimization. They were actively involved in addressing and resolving bugs related to line number handling and debugging. Moreover, the user implemented and improved features related to various parts of the optimizer framework to enable new functionality, as evidenced by their involvement in handling the internals of torch function mode. Additionally, the user worked on ensuring the accurate reconstruction of tensor subclasses with the framework, demonstrating a deep understanding of PyTorch's inner workings and the complexities of deep learning frameworks.
gpu-accelerationneural-networkpythonautogradgpu
mlazos/mlazos.github.io
Jun 2017 - Sep 2021
Contributions:52 pushes, 7 branches in 4 years 3 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.