Tracy Sharpe
Senior Staff Engineer at Modular
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolTracy Sharpe is a Senior Staff Engineer based in Seattle with seven years of recent experience and a long history of senior engineering roles, including Principal and Distinguished Engineer positions at Unity and Microsoft. She specializes in backend performance engineering, having contributed optimizations to the high-profile microsoft/onnxruntime project—improving MLAS matrix routines, adding an ARM64 U8S8 GEMM kernel, and refining 2D convolution kernels to accelerate inference. At Modular she continues to focus on scalable, high-performance systems while bringing deep expertise in low-level numeric kernels and cross-platform build compatibility. Tracy pairs hands-on performance tuning with architectural leadership, mentoring teams to ship robust, production-grade ML and systems software. Her career path reflects sustained technical impact across large platform teams, and her open-source contributions underscore a pragmatic focus on measurable performance gains.
7 years of coding experience
29 years of employment as a software developer
B.S. Software Engineer Technology, B.S. Software Engineer Technology at Oregon Institute of Technology
Github Skills (12)
quantization10
arm10
c-language10
cprogramming-language10
performance-optimization10
linear-algebra10
neural-network9
machine-learning9
convolutional-neural-networks9
deep-learning9
cmake8
build-system8
Programming languages (2)
Github contributions (5)
microsoft/onnxruntime
Nov 2018 - Oct 2021
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:188 reviews, 137 commits, 222 PRs in 2 years 11 months
Contributions summary:Tracy made several commits focused on optimizing the performance of core matrix operations within the ONNX Runtime project, specifically the MLAS (Microsoft Linear Algebra Subprograms) library. These optimizations included improvements to existing routines, such as the addition of a U8S8 GEMM kernel for ARM64, the development of improved 2D convolution kernels, and refactoring of several routines to streamline performance. Furthermore, the user addressed build issues and deprecated methods to maintain compatibility with various build environments.
runtimetrainingtensorflowai-frameworkaccelerator
yufenglee/onnxruntime
Apr 2019 - May 2019
ONNX Runtime: cross-platform, high performance scoring engine for ML models
Contributions:42 commits in 1 month
pytorchml-modelsscoringdeep-learningruntime
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.