About

I am Tom,

A Data Scientist at SeamlessML, utilizing Machine Learning to forecast alternative financial markets. In 2020, I graduated with an MEng from the University of Cambridge, specializing in Information Engineering and Computational Neuroscience.

My interests involve developing principled Bayesian models. Specifically, models that are well-calibrated and ‘know what they don’t know’, and models which are able to inherently deal with missing features and scarce/unlabelled data. I am also keen on investigating how model performance can be boosted to state-of-the-art using innovative feature-engineering and ensembling methods.

My Master’s thesis titled Deep Learning for Koopman Operator Optimal Control, utilises mathematically grounded and interpretable neural network architectures for optimally controlling non-linear dynamics. Previously I held internship positions at ARM as a Data Engineer and Onfido as a Machine Learning Research Engineer.