Frank Dong
Solutions Architect at Alberta Health Services
Edmonton, Alberta, Canada
Summary
🤩
Rockstar🎓
Top SchoolFrank Dong is a solutions architect and senior full-stack developer based in Edmonton with a multi-decade career delivering enterprise systems across government, healthcare, education, and industrial automation. He blends hands-on .NET and full-stack development experience with architecture and modernization leadership for large public-sector programs, currently shaping solutions at Alberta Health Services. Frank’s background includes deep work on identity & access management, enterprise security, and data exchange platforms, and he brings practical DevOps and agile delivery discipline to complex, regulated environments. He also contributes to high-profile open-source ML infrastructure—helping scale ONNX Runtime for Llama-70b through model fusion, sharding and BF16 support—showing an uncommon cross-over between enterprise architecture and cutting-edge machine learning performance engineering.
3 years of coding experience
24 years of employment as a software developer
Master of Engineering (M.Eng.), Computer Engineering, Master of Engineering (M.Eng.), Computer Engineering at University of AlbertaBachelor of Applied Science - BASc, Computer Science, Bachelor of Applied Science - BASc, Computer Science at Air Force Engineering University
Github Skills (11)
machine-learning10
deep-learning10
onnx10
hardware-acceleration10
python10
pytorch9
c-language8
cuda8
cprogramming-language8
tensorflow7
ai-framework7
Programming languages (2)
Github contributions (5)
microsoft/onnxruntime
Apr 2023 - Oct 2024
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:
ML Engineer
ML EngineerContributions:56 reviews, 17 PRs, 75 pushes in 1 year 6 months
Contributions summary:Frank focused on enhancing ONNX Runtime's capabilities for large language models, particularly the Llama-70b model. Their contributions included model fusion and sharding techniques to handle the model's size, alongside adding support for BF16 data type for various operations like Exp, Erf, and Convolution. They also addressed bugs related to distributed conversion and benchmarking for Llama2-70b and enabled the serialization of prepacked weights to reduce memory usage.
runtimetrainingtensorflowai-frameworkaccelerator
frank-dong-ms/onnxruntime
Oct 2023 - Jul 2024
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Contributions:22 pushes, 4 branches in 8 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.