Suffian Khan
Senior Software Engineer at Google
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolSuffian Khan is a Senior Software Engineer in Seattle with six years of industry experience building large-scale AI and cloud infrastructure, currently contributing to Google's GCP and AI supercomputer efforts. He brings deep HPC and performance-computing expertise from roles at Intel and Oak Ridge, where he architected distributed computation platforms, optimized physics and transistor modeling codes, and ran production-scale Monte Carlo jobs across tens of thousands of cores. At Microsoft and in open source he worked on AI inferencing and training frameworks—contributing CUDA/ROCm kernel fixes and training-resumption reliability to the widely used ONNX Runtime. A PhD-trained computational physicist with top grades in math and physics, he blends rigorous numerical modeling with practical API and tooling design (Python front-ends with C++ back-ends). Colleagues rely on him for diagnosing elusive parallel faults and squeezing order-of-magnitude performance gains from legacy codes. He is a U.S. citizen comfortable spanning research-grade simulation and production ML infrastructure.
6 years of coding experience
7 years of employment as a software developer
University of Illinois Urbana-ChampaignBachelor of Science (BS), Mathematics and Physics, 3.93 of 4.00, Bachelor of Science (BS), Mathematics and Physics, 3.93 of 4.00 at University of Florida
Github Skills (11)
cuda10
pytorch10
machine-learning10
deeplearning-ai10
roc10
deep-learning10
onnx9
hardware-acceleration9
c-language8
python8
cprogramming-language8
Programming languages (5)
Github contributions (5)
microsoft/onnxruntime
Apr 2020 - Dec 2021
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:
ML Engineer
ML EngineerContributions:158 reviews, 79 commits, 126 PRs in 1 year 8 months
Contributions summary:Suffian's commits primarily focus on enhancing the ONNX Runtime's training capabilities. Their work includes fixing checkpoint resumption in the `ort_trainer.py` script and adding unit tests. The user also implemented and debugged several CUDA kernels related to softmax operations and gradient calculations. Furthermore, the user optimized the code for the ROCm platform.
runtimetrainingtensorflowai-frameworkaccelerator
suffiank/ops-agent
Aug 2022 - Jan 2023
Contributions:108 pushes, 5 branches in 4 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.