Hi, and thanks for visiting my page. If you are interested in my public research, you can find a selected list of my work here, or on my Google Scholar profile.

I am excited by AI for Science and the prospects of “autonomous” scientific discovery: systems that tighten the loop between hypothesis, experiment, and insight, not replacing scientists and engineers but allowing them to work at ever higher levels of abstraction.

I am currently at Recursion / Valence Labs, where we learn models of biology from large-scale perturbation data to accelerate and improve drug discovery. My current focus is on bridging the scarce but highly informative signal from patient data with the massive scale of controlled in-vitro experiments, so that hypotheses grounded in patient biology can be exhaustively explored in-silico. I joined Recursion through the acquisition of Exscientia, where I led ML research in precision medicine.

Before that, I spent four years at the autonomous-driving startup FiveAI (acquired by Bosch), where I worked on ML and computer vision for robust and adaptive perception, and I had the fortune to mentor several talented PhD students.

I completed my PhD at the University of Oxford (2014–2018) during the exhilarating early days of deep learning, where I published some of the earliest work on deep few-shot and meta-learning, and developed the fully-convolutional Siamese networks tracking framework, which a broad family of subsequent methods built on.