Arthur Mitrano
Software Engineer at NVIDIA
Oregon, United States
Summary
🤩
Rockstar🎓
Top SchoolArthur Mitrano is a software engineer with 7 years of industry experience combining deep academic training (PhD in Applied Mathematics) with hands-on performance engineering at Intel and NVIDIA. He specializes in numerical analysis, computational fluid dynamics, scientific computing and high-performance optimizations, notably contributing performance-critical kernels to the widely used oneDNN library for AVX2, VNNI and bfloat16 workloads. Comfortable in both research and production settings, he has published in scientific journals, presented at international conferences, and taught calculus and computational methods. Fluent in English and Portuguese and based in Oregon, he brings a rare blend of rigorous mathematical foundations and low-level optimization expertise that accelerates ML and HPC workloads.
7 years of coding experience
14 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Applied Mathematics, Doctor of Philosophy (Ph.D.) Applied Mathematics at Arizona State UniversityMaster's Degree Applied Mathematics, Master's Degree Applied Mathematics at Universidade Federal do Rio de Janeiro
English, Portuguese
Github Skills (8)
avx10
matrix-multiplication10
c-language10
deep-learning10
cprogramming-language10
performance-optimization10
x86-6410
x8610
Programming languages (1)
Github contributions (4)
uxlfoundation/oneDNN
Nov 2018 - Jan 2023
oneAPI Deep Neural Network Library (oneDNN)
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:14 reviews, 177 commits, 44 comments in 4 years 2 months
Contributions summary:Arthur focused on optimizing the performance of the oneAPI Deep Neural Network Library (oneDNN). Their contributions centered around implementing and refining compute kernels for the s8x8s32 matrix multiplication operation, specifically for AVX2 and VNNI-enabled architectures. They implemented a driver quick exit, threading optimizations, and improved cache usage to enhance performance. Furthermore, the user added kernels for s8x8u32 and bfloat16 matrices.
bfloat16sse41avx512openmpopencl
openvinotoolkit/oneDNN
Oct 2019 - Aug 2020
oneAPI Deep Neural Network Library (oneDNN)
Contributions:6 commits in 10 months
caffe2deep-learningdeep-neural-networkoneapineural-network
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.