BiographyPhD student with professor Peter Grünwald at the Machine Learning group at CWI, and professor Floortje Scheepers at the Brain Center at UMC Utrecht. With a background in statistics and medicine, I like to work on bridging the gap between theoretical and applied statistics and to bring new theoretical concepts into practice.
With Safe Statistics, we develop more flexible and dynamical methods for statistical analysis. For example, safe statistics enable combining results of multiple experiments, even when this is done in hindsight, without losing type-I error control. Safe statistics also enable optional stopping of experiments, potentially saving resources through early stopping. I am working on developing "safe" versions of various classical hypothesis tests and implementing these tests in software packages.
Enabling Personalized Interventions
Within the "Enabling Personalized Interventions" project we aim to enable more personalized interventions or recommendations for patients in healthcare, through combining different sources of eHealth data, and liberating and analyzing that data in a trustworthy way. I work on a use case where we aim to create explainable recommendations for patients in psychiatry, with statistically valid estimations of associations between patient characteristics, interventions and outcomes based on combined models from various mental health hubs.