Wonjoo Lee - Software Engineer at Meta

Wonjoo Lee

Software Engineer at Meta

San Francisco, California, United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Wonjoo Lee is a software engineer with 10 years of experience building scalable ML infrastructure and backend platforms across Google, Uber, AWS, and Meta. He specializes in ML frameworks and inference for large models, notably optimizing PyTorch/XLA for TPU performance and integrating with vLLM. His open-source contributions to PyTorch include lazy shape inference and XLA lowerings for operations like slogdet and mish, showing deep backend and ML systems expertise. At Google and WaveForms AI he focused on productionizing LLM inference; at Uber and AWS he built resilient large-scale backend services and automation. Based in San Francisco, he blends research-grade contributions with production engineering, often working at the intersection of compiler-level optimizations and deployment systems. Colleagues describe him as a pragmatic problem-solver who moves complex ML systems from prototype to production.

10 years of coding experience

7 years of employment as a software developer

University of California, Irvine

Stackoverflow

Stats

111reputation

9kreached

1answer

7questions

Github Skills (25)

pytorch10

python10

inference10

ml10

mle10

deep-learning10

xla10

web-framework10

tensor10

shapes10

c-language9

machine-learning9

cprogramming-language9

compiler8

compiler-compiler8

Programming languages (6)

JavaC++ShellRubyPythonJsonnet

Github contributions (5)

pytorch/xla

Oct 2021 - Jan 2023

Enabling PyTorch on XLA Devices (e.g. Google TPU)

Role in this project:

Back-end Developer & ML Engineer

Contributions:1 release, 504 reviews, 578 commits in 1 year 3 months

Contributions summary:Wonjoo primarily contributed to enabling PyTorch on XLA devices. They implemented operation lowerings for various functions, including `slogdet`, `prelu`, and `mish`. Furthermore, the user added support for these functions by creating unit tests. Additionally, the user made updates to the existing code base including fixes to linter issues and adapting to new PyTorch APIs.

pytorchxladeep-learningtpucompiler

pytorch/pytorch

Jan 2022 - Jan 2023

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Role in this project:

Back-end Developer

Contributions:15 reviews, 25 commits, 40 PRs in 1 year

Contributions summary:Wonjoo primarily focused on adding and modifying lazy shape inference functions within the PyTorch framework. They implemented shape inference for logical boolean operations, including `logical_and`, `logical_not`, `logical_or`, and `logical_xor`, improving the efficiency of shape determination in the library. They also added shape inference for the `take` and `cholesky` operations. Furthermore, the user updated the codegen to customize the GenLazyNativeFuncDefinition generator for future XLA integration.

pythongpu-accelerationdeep-learninggpunumpy

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.

Request Free Trial