Lily Orth-smith
Software Engineer at Microsoft
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolLily Orth-smith is a machine learning compilers engineer with a decade of experience building production-quality compiler infrastructure for heterogeneous ML hardware, currently working on the MAIA chip at Microsoft. She has deep expertise in programming languages, type theory, and dynamic-shape support—evidenced by significant contributions to the popular Apache TVM project, where she implemented dynamic operators and refined dynamic-to-static passes. Prior roles at Modular AI and OctoML focused on unifying compilers and runtimes for edge-to-datacenter workflows, and she’s comfortable changing core lowering and pass architectures to improve analyzability. Based in Seattle, she pairs rigorous systems thinking with pragmatic engineering and a curiosity for formal methods. Outside of work she’s an avid hiker and salsa dancer, with hobbies that reflect a hands-on, exploratory mindset from ceramics to urban foraging.
10 years of coding experience
4 years of employment as a software developer
Bachelor of Arts - BA Computer Science, Bachelor of Arts - BA Computer Science at Wellesley College
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (9)
compiler10
tvm10
machine-learning10
deeplearning-ai10
compiler-compiler10
c-language10
deep-learning10
cprogramming-language10
python9
Programming languages (7)
Github contributions (5)
apache/tvm
Jul 2020 - Feb 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:162 reviews, 38 commits, 44 PRs in 1 year 7 months
Contributions summary:Lily primarily contributed to implementing and refining dynamic operator support within the TVM compiler stack. Their work focused on adding new dynamic operators, specifically for broadcast_to, zeros, ones, full, and pad, and included developing corresponding type relations, compute functions, and registration. Additionally, the user implemented the dynamic full operator, tests for these dynamic operators, and also worked on refining the dynamic to static pass. The user also worked on dynamic upsampling operators and implementing pad operator.
metalvulkancompilertensoropencl
electriclilies/tvm
Jun 2020 - Feb 2022
Open deep learning compiler stack for cpu, gpu and specialized accelerators
Contributions:10 reviews, 5 PRs, 673 pushes in 1 year 8 months
cpugpu-programminggpu-accelerationtvmdeep-learning
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.