Tom Tseng
Research Engineer at FAR.AI
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolTom Tseng is a research engineer in the San Francisco Bay Area with 11 years of software and ML experience focused on reducing future AI risks and improving robustness. At FAR.AI he led practical attacks against superhuman Go AIs and explored defenses, showing that capabilities don't imply robustness and that simple adversarial training can fall short. He blends research-grade algorithm development from MIT and CMU training with product-focused engineering at startups like Gather Town and Cruise, shipping performance and reliability improvements in real systems. An active contributor to open-source ML tooling, he improved CI/CD and cross-platform testing for imitation learning codebases, removing heavy dependencies to ease adoption. Colleagues describe him as equally comfortable optimizing routing graphs and debugging streaming stacks, with a curiosity for surprising failure modes in deployed AI.
11 years of coding experience
4 years of employment as a software developer
Master of Science - SM Electrical Engineering and Computer Science, Master of Science - SM Electrical Engineering and Computer Science at Massachusetts Institute of TechnologyHigh schoolBachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at Carnegie Mellon University
English, Chinese
Stackoverflow
155reputation
450reached
2answers
0questions
Github Skills (14)
machine-learning10
python10
cicd10
githubaction-workflow9
github-ci9
imitation-learning9
gymnasium8
sphinx7
testing7
documentation7
pytorch7
sml6
functional-programming6
set-theory6
Programming languages (11)
Github contributions (5)
HumanCompatibleAI/imitation
Aug 2022 - Aug 2022
Clean PyTorch implementations of imitation and reward learning algorithms
Role in this project:
DevOps Engineer & ML Engineer
DevOps Engineer & ML EngineerContributions:1 release, 18 reviews, 8 commits in 7 days
Contributions summary:Tom primarily focused on improving the continuous integration and continuous deployment (CI/CD) pipeline by adding macOS tests to CircleCI and removing a MuJoCo requirement. They also made updates to the documentation by integrating release notes. Furthermore, the user made changes to the code related to preference comparisons, including saving untransposed trajectories.
pytorchimplementationsreinforcement-learningcleanmachine-learning
tomtseng/parallel-euler-tour-tree
Dec 2018 - Aug 2022
Code for "Batch-Parallel Euler Tour Trees" paper
Contributions:1 release, 2 reviews, 2 commits in 3 years 8 months
eulerparalleltreesbatch
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.