Frost Mitchell
AI Frameworks Engineer at Intel Corporation
Washington DC-Baltimore Area United States
Summary
🤩
Rockstar🎓
Top SchoolFrost Mitchell is an AI Frameworks Engineer with three years of hands-on experience building and optimizing deep learning infrastructure, currently driving PyTorch-related feature work at Intel in the Washington DC–Baltimore area. His contributions to the flagship pytorch/pytorch repository include improving torch.fx subgraph rewriting and adding a C++ kernel template for batched matrix multiplication in the inductor backend, addressing precision support (FP32/FP16/BF16) and dynamic batching—work that improves both correctness and performance for GPU workloads. He combines academic depth (PhD-level computer science training at the University of Utah) with practical systems experience from internships and embedded/defense projects, ranging from RF signal synthesis to real-time image classification. Frost’s background in adversarial ML and low-level kernel work gives him a rare blend of algorithmic insight and systems-level optimization skills, making him effective at shipping production-grade ML framework enhancements.
3 years of coding experience
3 years of employment as a software developer
Bachelor of Science - BS, Computer Engineering, Bachelor of Science - BS, Computer Engineering at Utah State UniversityThe University of Utah
Github Skills (11)
pytorch10
machine-learning10
c-language10
cprogramming-language10
python10
deep-learning9
autograd8
gpu8
numpy7
bfd7
tensor7
Programming languages (2)
Github contributions (5)
pytorch/pytorch
Mar 2023 - Mar 2025
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Role in this project:
ML Engineer
ML EngineerContributions:53 reviews, 6 PRs, 96 comments in 1 year 11 months
Contributions summary:Frost contributed to the `pytorch/pytorch` repository, a project focused on tensors and dynamic neural networks. Their work involved enhancing the `torch.fx.subgraph_rewriter` module, specifically addressing issues related to recursion and updating internal data structures after repeated replacements, leading to improved performance and correctness. Additionally, the user implemented a Cpp kernel template for BMM operations within the inductor framework, adding support for FP32, FP16, and BF16 data types and resolving issues with dynamic batch sizes and FlexibleLayout weights. This suggests a focus on optimizing and extending the capabilities of PyTorch's backend for efficient matrix multiplication.
pythongpu-accelerationdeep-learninggpunumpy
frost-intel/pytorch
Mar 2023 - Mar 2025
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:123 pushes, 12 branches in 1 year 11 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.