About

About me

I am a Senior Researcher in Machine Learning at Microsoft Research Cambridge. Currently, I am primarily interested in (re)building statistical intution for modern ML methods and apparently elusive phenomena in deep learning – such as double descent and the surprisingly good performance of models that would appear to be heavily overfitted – as I strongly believe in the importance of demystification of the area. Most of my earlier research centered around how to best use ML to estimate personalized causal effects of treatments. Overall, I am broadly interested in many topics relating to causality, missing data, distribution shifts and (essentially) all of statistical machine learning. You can find a selection of my work under the Publications page or on my Google Scholar Profile. Please feel free to reach out in case you want to discuss with me!

My professional background

Albeit a ML researcher today, I am a statistician at heart – and most of my research therefore reflects my desire to connect ideas from (applied) statistics with machine learning. I hold a BSc in Econometrics and Operations Research (summa cum laude) and a BSc in Economics and Business Economics (summa cum laude) with specialisation in Policy Economics from the Erasmus University Rotterdam, and a MSc in Statistical Science (with distinction and prize for top performance in the cohort) from the University of Oxford. Between 2020 and 2024, I completed a PhD in Machine Learning at the University of Cambridge, advised by Prof. Mihaela van der Schaar. Before starting the PhD, I also briefly worked as a data scientist evaluating the effectiveness of marketing campaigns, and as a research intern at Pacmed, a Dutch start-up bringing ML solutions to clinical practice.

Me, personally

When not thinking about stats and ML, I like to stay active. I played waterpolo for 14 years (I earned half-blue status in both Oxford and Cambridge by representing both universities in some of the annual varsity matches) and – like every other Oxbridge student – became infatuated with rowing during my student time. One may sometimes find me away from water on a run or a hike. I am also always ready to binge-watch a good Netflix show and love good food & cooking.

News!

• Nov ‘24: The last paper of my PhD, on a simple model for understanding modern ML phenomena, was accepted to NeurIPS24. Find it here!
• Oct ‘24: Big life update – I’ve handed in my PhD thesis and joined Microsoft Research Cambridge as a Senior Researcher in Machine Learning!
• Sep ‘24: Wrote a little note on what I wish I knew as a Masters student to understand modern ML phenomena like double descent, read it here!
• Feb ‘24: New preprint ‘Why do Random Forests Work?’ (check it out!)
• Jan ‘24: Our paper ‘Cautionary Tales On Synthetic Controls in Survival Analyses’ was accepted as an Oral Presentation at the Conference on Causal Learning and Reasoning (CLeaR) happening in April 2024!
• Jan ‘24: Our review ‘Using Machine Learning to Individualize Treatment Effect Estimation: Challenges and Opportunities’ was accepted for publication in Clinical Pharmacology & Therapeutics, check it out here!
• Dec ‘23: I was interviewed on the ‘Causal Bandits’ Podcast, check it out here!
• Sep ‘23: Our paper ‘A U-turn on Double Descent’ (joint work with Alan Jeffares) has been accepted to NeurIPS23 as an Oral!!
• Aug ‘23: I will be interning with Javier González in the Health Futures team of Microsoft Research for the next 3 months (August-October)!
• Jul ‘23: I am going to Honolulu and I am bringing… three posters, on topics relating to treatment effect estimation, to ICML23!
• April ‘23: Find me in Valencia presenting a poster on heterogeneous treatment effect estimation in the presence of competing risks at AISTATS23!
• Dec ‘22: Check out our paper on Benchmarking Heterogeneous Treatment Effect Estimators through the Lens of Interpretability at NeurIPS22!
• Jul ‘22: I will be attending my first in-person conference, ICML22, in Baltimore – presenting work on Automated Imputation at the main conference, and work on Adaptive Clinical Trials at a workshop!
• Dec ‘21: 3 papers accepted at NeurIPS21 – including a spotlight on inductive biases in heterogeneous treatment effect estimation!
• April ‘21: I will be presenting the first paper of my PhD – on heterogeneous treatment effect estimation – at AISTATS21!
• Oct ‘20: I’ve started a PhD with Prof. Mihaela van der Schaar at the Department of Applied Mathematics at the University of Cambridge!