Welcome! I’m Felix Zimmer, a Postdoctoral Researcher at the Singapore-ETH Centre (ETH Zurich), where I work in Prof. Torsten Hoefler’s FastTrackAI programme. I received my doctoral degree from the University of Zurich, where I specialized in applying machine learning to statistical methodology under the supervision of Dr. Rudolf Debelak.
My research has moved from quantitative psychology toward efficient and trustworthy machine learning. At the moment I work on SciTorch, which helps AI agents find the right methods when they build scientific models, and on medical-image foundation models trained on the Alps supercomputer. Before that, I developed EntryPrune, a feature selection method for neural networks based on dynamic sparsity.
Outside of research, I enjoy cycling 🚴, bouldering 🧗, and eating lots of Korean ramen 🍜.
Recent updates
- 06/2026. I joined the Singapore-ETH Centre as a Postdoctoral Researcher.
- 03/2026. I started as a Scientific Assistant at the Scalable Parallel Computing Laboratory at ETH Zurich.
- 05/2025. I’ve teamed up with great collaborators for my project on using sparsity for feature selection, which has resulted in a significant update to the preprint.
- 01/2025. We received a research award to develop software for sample size calculations for clinical prediction models.
- 10/2024. My paper exploring sparsity-based feature selection in neural networks is now available as a preprint.
- 12/2023. We developed a new framework for optimizing complex study designs to achieve desired statistical power at minimal costs. Our paper is published in Psychological Methods.
- 11/2023. We provide a tutorial for the
mlpwrpackage to optimize complex study designs (Behavior Research Methods, CRAN). - 10/2023. We propose new formulas for calculating the necessary sample size when using IRT models. Our paper is published in Psychometrika and our package
irtpwris available on CRAN. - 04/2023. Successfully defended my PhD thesis in which I developed new methods for power analysis and sample size planning.