Nichols A. Romero
Naperville, Illinois, United States
Summary
🤩
RockstarNichols Romero is a computational scientist and software engineer who brings two decades of domain expertise to enable high-performance ML and scientific workloads, currently focused on making PyTorch and PyTorch-based models run efficiently on AMD GPUs. His career spans national labs and industry—Argonne, Intel, and Samsung—where he led performance engineering, compiler front-end work for OpenMP on GPUs, and co-designed next-generation supercomputing architectures. He has deep hands-on experience optimizing scientific codes (QMCPACK, GPAW, MADNESS) for leadership-class systems and contributed tooling like an HPC-centric Python installer and an OpenACC-to-OpenMP converter for Fortran. Based in Naperville, IL, Nichols combines PhD-level physics training with practical systems and compiler skills, mentoring teams and shaping strategic plans for exascale projects. Notably, he bridges vendor, hardware, and open-source ecosystems—having worked directly with Intel math libraries and now addressing community issues on AMD’s PyTorch team.
2 years of coding experience
Github Skills (59)
pytorch10
python10
windows10
llama10
acceleration10
machine-learning10
linux10
inference10
numpy10
amd10
llm10
deep-learning10
tensorflow10
gpu10
autograd10
Programming languages (5)
Github contributions (5)
naromero77amd/tutorials
Dec 2024 - Mar 2025
PyTorch tutorials.
Contributions:5 pushes, 2 branches in 2 months
ROCm/pytorch
Jul 2024 - Apr 2025
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:1 review, 9 PRs, 130 pushes in 9 months
pythongpu-accelerationdeep-learninggpuacceleration
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.