Adrian Tsai
Software Engineer at Meta
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolAdrian Tsai is a software engineer with eight years of experience specializing in high-performance graphics and machine learning acceleration, now based in Redmond and currently at Meta. He spent a decade at Microsoft advancing DirectML and DirectX integration—working from Visual Studio graphics diagnostics to leading DirectML efforts—and contributed key graph transformations and execution-provider support to the widely used onnxruntime project. Adrian’s strengths lie at the intersection of low-level execution engines and ML inference optimization, including implementing a FreeDimensionOverrideTransformer and DirectML operator support that improved hardware-accelerated ML pipelines. He pairs hands-on coding and testing discipline with a track record of shipping platform-level improvements that surface as better performance and broader hardware compatibility.
7 years of coding experience
12 years of employment as a software developer
The University of Melbourne
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (13)
machine-learning10
c-language10
deep-learning10
tensorflow10
onnx10
cprogramming-language10
directml10
directx9
hardware-acceleration9
cmake8
api8
apidoc8
pytorch7
Programming languages (3)
Github contributions (5)
microsoft/DirectML
Mar 2019 - Sep 2022
DirectML is a high-performance, hardware-accelerated DirectX 12 library for machine learning. DirectML provides GPU acceleration for common machine learning tasks across a broad range of supported hardware and drivers, including all DirectX 12-capable GPUs from vendors such as AMD, Intel, NVIDIA, and Qualcomm.
Role in this project:
ML Engineer
ML EngineerContributions:58 reviews, 15 commits, 41 PRs in 3 years 6 months
Contributions summary:Adrian primarily contributes to the development and enhancement of the DirectML library, a high-performance DirectX 12 library for machine learning. Their work includes initial implementation of sample code demonstrating DirectML usage, refactoring and updating existing code, fixing newline formatting issues. The user also added and maintained TensorFlow samples to the repository. Furthermore, they made improvements to the DMLX library, including the implementation of new operators and the ability to specify output sizes for existing operators.
amdaccelerationdirectmlgpu-accelerationnvidia
microsoft/onnxruntime
Aug 2019 - Jun 2022
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:Back-end Developer
Back-end DeveloperContributions:27 reviews, 28 commits, 12 PRs in 2 years 11 months
Contributions summary:Adrian's contributions focused on implementing and optimizing graph transformations within the ONNX Runtime. They developed a FreeDimensionOverrideTransformer, including its implementation, testing, and integration within the graph optimization pipeline. Furthermore, they added support for the DirectML execution provider, including associated operator registrations and fixes, demonstrating involvement in the execution engine and hardware acceleration aspects of the project. The user also addressed build and testing issues related to the DirectML execution provider.
runtimetrainingtensorflowai-frameworkaccelerator
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.