I build machine learning systems, and take them from research to production.

My main interests are in ML applications to health and life sciences, working on new methods in optimization, Bayesian experimental design and causal inference on clinical data.

BExTE — Bayesian Extrapolation of Treatment Effects: a clinical trial-design simulation tool that operationalizes my EMA-funded research on extrapolating treatment effects across populations. Docs.

statrl — contributor to the Statistical Reinforcement Learning Toolkit, an open-source library for statistical RL.

I've developed the AI platform SEARIDE for seahorse reidentification from photographs for citizen science programs.

Languages

Python · MATLAB · R

Bayesian & probabilistic ML

Stan · PyMC · GPflow · BoTorch

Machine learning

PyTorch · scikit-learn

MLOps

Docker · Git · CI/CD · MLflow · Google Cloud

Bayesian Optimization · Experimental Design · Causal Inference · Reinforcement Learning · Gaussian Processes · Human-in-the-Loop Optimization · Clinical Trial Design · Real-World Evidence