Dwayne Robinson
Software Development Engineer at Microsoft
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolDwayne Robinson is a seasoned Software Development Engineer with nearly two decades at Microsoft focused on graphics, text layout, and GPU-accelerated machine learning. He spent a decade shaping DirectWrite typography (line breaking, bidi, font selection and variable-font support) and then pivoted to DirectML, implementing numerous HLSL operators and extending the DML API to accelerate ONNX/WebNN workloads. Dwayne contributes to high-profile open source projects like ONNX Runtime (kernel work, optimizer fusions, DirectML provider fixes) and has improved core C++ graphics libraries such as lunasvg, demonstrating deep systems-level expertise. Based in Redmond, he blends low-level shader/kernel development with practical API design and spec work, having even co-edited WebNN operator specifications. Outside code he’s also an artist and pianist, a detail that reflects a longtime creative interest informing his approach to graphics and tooling.
10 years of coding experience
MS, Computer Science, MS, Computer Science at Oregon State University
Stackoverflow
2,249reputation
259kreached
45answers
0questions
Github Skills (31)
bitmap10
c-language10
bit-manipulation10
operation10
directml10
tensorrt10
machine-learning10
svg10
tensorflow10
onnxruntime10
imagebitmap10
bitmaps10
tensor10
cpp10
cprogramming-language10
Programming languages (20)
Github contributions (5)
microsoft/onnxruntime
Jul 2019 - Jan 2023
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1656 reviews, 210 commits, 225 PRs in 3 years 6 months
Contributions summary:Dwayne's contributions centered around optimizing the performance of the ONNX Runtime for machine learning inference. They implemented float16 MatMul+Add -> GEMM fusion to improve performance by modifying the optimizer. Additionally, they addressed a test failure in the DirectML provider's TensorDesc class, focusing on tensor placement tests. The user's work is related to implementing kernels and adding functionalities for the DirectML execution provider, demonstrating strong understanding of the ONNX Runtime codebase.
runtimetrainingtensorflowai-frameworkaccelerator
sammycage/lunasvg
Feb 2022 - May 2022
SVG rendering and manipulation library in C++
Role in this project:Backend Developer
Backend DeveloperContributions:5 commits, 4 PRs, 8 comments in 2 months
Contributions summary:Dwayne primarily focused on fixing and improving the `Bitmap` class, which appears to be a core component for handling image data. Their contributions involved correcting stride calculations and data conversions within the `Bitmap::clear` and `Bitmap::convert` functions, addressing potential issues. Additionally, the user made consistent changes to improve code readability by using `std::uint32_t` and fixed an issue related to href parsing. They also updated the default value for stop offset within the code.
cppsvgstandalonerenderingscenegraph
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.