Adam Roberts
Director Of Research at Google DeepMind
California, United States
Summary
🤩
Rockstar🎓
Top SchoolAdam Roberts is a Director of Research at Google DeepMind with 15 years of experience building foundational AI systems and leading teams that bridge research and production. He has driven infrastructure and model development across high-profile projects at Google—T5, PaLM, MusicLM and more—while contributing substantial open-source improvements to libraries like JAX, Flax, TensorFlow.js and T5X. Equally comfortable in ML research and engineering, he has improved model training, checkpointing, dataset pipelines and concurrency in both Python and JavaScript ecosystems. His background in music and art generation (Magenta, MT3) and hands-on work on quantized weights and WebGL-accelerated TFJS show a rare combination of systems-level rigor and creative ML applications. A PhD-trained computer scientist, he also brings prior domain breadth from genomics tooling to product signals teams and even classroom teaching, which informs his pragmatic mentorship and cross-disciplinary perspective.
15 years of coding experience
17 years of employment as a software developer
PhD Computer Science, PhD Computer Science at University of California, BerkeleyBS Computer Science Mathematics Physics, BS Computer Science Mathematics Physics at University of North Carolina at Chapel Hill
Spanish
Github Skills (99)
python10
multithreading10
testing10
refactor10
flax10
datasets10
transformer-models10
deep-learning10
trainings10
refactoring10
threaded10
concurrency10
javascript10
protocol-buffers10
tensorflowjs10
Programming languages (8)
Github contributions (5)
google-research/t5x
Nov 2021 - Jan 2023
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:2 reviews, 149 commits, 4 PRs in 1 year 2 months
Contributions summary:Adam primarily focused on improving the project's codebase and infrastructure. Their contributions involved moving and updating Git workflows, and fixing tests. Furthermore, the user improved the installation process for the project and made changes to the code to enable easier access to files within the installed library. They also worked on resolving Keras-related issues.
magenta/magenta-demos
Jun 2017 - Jul 2020
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:2 reviews, 28 commits, 36 PRs in 3 years 1 month
Contributions summary:Adam primarily worked on implementing the ai-jam-js demo, adding and modifying various components. Their contributions included adding the core ai-jam-js demo, integrating a notification bar, and adding controls. They also made updates to the README, fixed quantization errors, and added a missing script. Further, they made changes to various JavaScript files.
magentamachine-learningdemonstrations
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.