David Eriksson
Research Scientist at facebook
Boulder, Colorado, United States
Summary
🤩
RockstarDavid Eriksson is a Research Scientist and manager at Meta with 11 years of experience designing and shipping high-performance ML systems from Boulder, Colorado. He specializes in Gaussian processes and Bayesian optimization, contributing GPU acceleration, tests, and documentation to prominent open-source projects like gpytorch, BoTorch, and Ax. His work spans backend development and algorithmic fixes—ranging from acquisition-function bug fixes to implementing SAASBO and TuRBO tutorials—demonstrating both research depth and production rigor. Colleagues rely on him to bridge cutting-edge probabilistic modeling with practical, scalable implementations that run efficiently on CUDA-enabled hardware.
10 years of coding experience
Github Skills (11)
cuda10
pytorch10
machine-learning10
bayesian10
gpu-acceleration10
python10
optimisation10
optimization10
data-analysis9
develop9
unit-testing9
Programming languages (3)
Github contributions (5)
facebook/Ax
Aug 2020 - Oct 2022
Adaptive Experimentation Platform
Role in this project:
Backend Developer
Backend DeveloperContributions:32 commits, 70 PRs, 3 pushes in 2 years 2 months
Contributions summary:David implemented a new method `_get_candidate_metadata(arm_name)` across multiple files related to the `ax` experimentation platform, including `base_trial`, `trial`, and `batch_trial`. This involved modifying existing test files to incorporate and validate the new method. The changes also touched the `observation.py` file, indicating the integration of the new method within the core functionality of the experimentation platform.
experimentationadaptivesimulationadaptive-experimentation-platformexperimentation-platform
cornellius-gp/gpytorch
Dec 2018 - Jun 2022
A highly efficient implementation of Gaussian Processes in PyTorch
Role in this project:ML Engineer
ML EngineerContributions:20 reviews, 20 commits, 20 PRs in 3 years 6 months
Contributions summary:David contributed to the repository by adding CUDA tests, indicating a focus on GPU acceleration within the Gaussian Process framework. The user also added comments and documentation to example notebooks and implemented GPU support within the codebase. These changes suggest involvement in improving the performance and usability of Gaussian Process implementations in PyTorch.
pytorchgpu-accelerationgaussianstochastic-processesmodular
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.