About

Hi, I’m Thomas, a Senior Data Scientist in the Content and Studios team at Netflix. My work focusses on building causal models in domains where traditional AB testing is infeasible and we must, therefore, lean upon alternative tooling e.g., proximal causal inference, instrumental variables, and unobserved confounding. Often this requires applying Bayesian techniques. Prior to Netflix, I worked as a Senior Applied Scientist in the Maps team of Uber and Amazon’s Prime and marketing team.

Outside of work, I lead the development of GPJax, a Gaussian process package written in JAX, and Impulso, a framework for fitting Bayesian structural vector autoregression models.

Previously, I completed a PhD in Statistics at Lancaster University. My thesis was titled Developments in Gaussian processes with applications to networks and climate sciences.