Maximilian Balandat
Senior Research Scientist Manager at Meta
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMaximilian Balandat is a Senior Research Scientist Manager at Meta with 11 years of experience building and shipping Bayesian optimization and probabilistic modeling tools in production. He leads the team maintaining BoTorch and drives Modeling & Optimization for the Adaptive Experimentation group, combining deep PyTorch expertise with practical algorithm engineering. A UC Berkeley PhD in EECS, he blends control theory, economics, and statistics to push research-grade methods into scalable systems, and has contributed performance and kernel-level improvements to widely used projects like gpytorch and SciPy. His open-source work includes enhancing quasi-Monte Carlo sampling and implementing batched L-BFGS updates and ARD kernels, reflecting both math depth and backend systems skill. Based in San Francisco, he pairs manager‑level leadership with hands-on coding and testing framework design. Outside work he’s a certified PADI SCUBA instructor and scientific diver, an uncommon combination of technical rigor and field exploration.
11 years of coding experience
11 years of employment as a software developer
Master of Arts (M.A.) Mathematics, Master of Arts (M.A.) Mathematics at University of California, BerkeleyTechnischen Universität Darmstadt
German, English, French
Github Skills (25)
algorithm10
algorithms10
pytorch10
scipy10
python10
testing10
machine-learning10
machine-learning-algorithms10
numpy10
gaussian-processes10
kernel10
bayesian10
optimisation10
kernel-mode10
gpytorch10
Programming languages (6)
Github contributions (5)
pytorch/botorch
Oct 2018 - Jan 2023
Bayesian optimization in PyTorch
Role in this project:
ML Engineer & Back-end Developer
ML Engineer & Back-end DeveloperContributions:23 releases, 517 reviews, 642 commits in 4 years 3 months
Contributions summary:Maximilian implemented and optimized a batched version of L-BFGS updates, a core component of the QP solver used in the project. They also developed functions for evaluating model performance. The code modifications indicate the user was involved in integrating, testing, and performance improvements to existing functionality. Additionally, they were responsible for creating and integrating a testing framework into the project.
pytorchoptimizationmultiobjective-optimizationmachine-learningbayesian-optimization
facebook/Ax
May 2019 - Jan 2023
Adaptive Experimentation Platform
Role in this project:Back-end Developer
Back-end DeveloperContributions:1 release, 63 reviews, 107 commits in 3 years 8 months
Contributions summary:Maximilian primarily contributed to the `ax` library, a platform for adaptive experimentation. Their work focused on modifying and enhancing the BoTorch models used within the library. They made changes to acquisition functions, optimizers, and model configurations, indicating expertise in optimizing and refining the core algorithms. Additionally, they worked on incorporating the handling of discrete parameters and adjusting the underlying models for improved performance and functionality.
experimentationadaptivesimulationadaptive-experimentation-platformexperimentation-platform
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.