Gus Smith
Researcher at YosysHQ
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolGus Smith is a researcher and engineer blending 11 years of experience across machine learning compilers, programming languages, formal methods, and hardware design. He completed a PhD at the University of Washington where he built compiler-backend automation and novel ML-to-accelerator tooling (Glenside, Lakeroad/Churchroad) and has continued that thread in industry and open-source contributions. Gus has practical experience extending the widely-used TVM compiler with custom datatypes and now contributes to YosysHQ’s open-source EDA tooling while pursuing independent research with university collaborators. He has prototyped AI agents for hardware verification, helped drive adoption of open-source hardware tooling at national labs, and was part of a startup acquisition by Cadence. Based in Seattle, he combines deep theory with hands-on systems engineering and an unusual focus on making compiler and hardware toolchains more automated and agentic.
11 years of coding experience
5 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of WashingtonMaster of Science - MS Computer Science and Engineering, Master of Science - MS Computer Science and Engineering at Penn State University
English
Stackoverflow
1reputation
2kreached
0answers
1question
Github Skills (11)
compiler10
tvm10
machine-learning10
deeplearning-ai10
compiler-compiler10
deep-learning10
datatypes10
python9
c-language8
cprogramming-language8
sharepoint6
Programming languages (24)
Github contributions (5)
apache/tvm
May 2019 - May 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
ML Engineer
ML EngineerContributions:7 reviews, 17 commits, 20 PRs in 3 years
Contributions summary:Gus primarily focused on implementing and extending custom datatypes within the TVM compiler stack. Their contributions included registering custom datatypes, defining lowering functions for operations (like Cast, Add, Mul), and integrating these datatypes into the compilation process. They also worked on enabling the use of custom datatypes from Python and contributing to a "Bring Your Own Datatypes" framework. This suggests a focus on extending TVM's capabilities for supporting different data representations in machine learning workloads.
metalvulkancompilertensoropencl
gussmith23/posit-bisection-viz
Apr 2019 - Jun 2019
Contributions:78 commits, 1 PR, 58 pushes in 1 month
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.