Barry Dong
AI Research Scientist at Meta
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolBarry Dong is an AI research scientist with 11 years of experience building production-grade models and infrastructure across premier labs and startups. He co-created foundational products at DeepMind (including early Bard/Gemini work and optimizers like Lion) and later led RL-focused research and tooling as a founding researcher at Augment. Barry has hands-on engineering chops—contributing to open-source projects from AutoDL tooling and PyTorch-based training automation to speech toolkit compatibility fixes—and has a track record of improving code quality, reproducibility, and CI/cluster workflows. He spent time on the engineering side as a Member of Technical Staff at OpenAI and now works at Meta, blending research depth with pragmatic system design. Trained in computer science (Beihang University, PhD work at University of Technology Sydney), he uniquely pairs optimizer and RL research with practical DevOps and back-end automation experience. An interest-driven explorer, he often surfaces small but high-impact improvements—like richer reprs and cross-version compatibility—that make large ML systems more maintainable.
11 years of coding experience
3 years of employment as a software developer
Bachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at Beihang UniversityDoctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of Technology Sydney
Github Skills (29)
pytorch10
python10
scripting10
package-development10
machine-learning10
speech-synthesis10
script10
deep-learning10
nas10
quantitative-finance10
sh10
speech-recognition10
shell10
neural-architecture-search10
debug9
Programming languages (5)
Github contributions (5)
D-X-Y/Awesome-AutoDL
Oct 2018 - Sep 2022
Automated Deep Learning: Neural Architecture Search Is Not the End (a curated list of AutoDL resources and an in-depth analysis)
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 review, 184 commits, 40 PRs in 4 years
Contributions summary:Barry appears to be primarily focused on the development of the `awesome-autodl` Python package, specifically concerning the analysis of papers related to Automated Deep Learning. The contributions involve reorganizing the package structure, setting up the build process for PyPI, and creating scripts to analyze and categorize papers related to different AutoDL sub-topics. They also made changes to the data structures by adding new fields to the class, and updated the code in response to data changes.
pytorchpythonnasdeep-learningneural-architecture-search
D-X-Y/AutoDL-Projects
Feb 2019 - Apr 2022
Automated deep learning algorithms implemented in PyTorch.
Role in this project:Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:530 commits, 7 PRs, 476 pushes in 3 years 3 months
Contributions summary:Barry primarily focused on updating and modifying shell scripts related to a deep learning project, automating deep learning algorithms implemented in PyTorch. They made changes to training scripts for CNN models, as well as submit and job scripts related to a cluster environment, indicating involvement with infrastructure and possibly CI/CD aspects. They also updated the data pre-processing steps. Their contributions suggest an interest in automating model training and related tasks.
pytorchnasdeep-learningneural-architecture-searchautodl
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.