I am a Research Scientist at Prior Labs, developing the next generation of tabular foundation models. My work focuses on scalable pretraining, novel architectures, and scaling laws, contributing to TabPFN-3's expansion to datasets with up to one million training rows and substantially faster training and inference.
From 09/2024 to 01/2026, I was a Senior AI Researcher in Aleph Alpha Research's Foundation Models team, where I focused on LLM pretraining and optimization. I contributed to tokenizer-free LLM architectures for efficient pretraining, domain adaptation, and inference, as well as optimization methods for efficient training of large models.
I was Senior Expert at Bosch Center for Artificial Intelligence (BCAI) until 08/2024. As a Senior Expert in Deep Learning at Bosch Center for Artificial Intelligence (BCAI), my primary research focused on making AI (specifically computer-vision based perception) robust, reliable, and safe. This included identifying vulnerabilities of Transformer-based neural network against adversarial patch/token attacks as well as finding systematic errors of image classifiers on rare subgroups and systematic errors of object detectors. We also developed architectures that are certifiably robust against patch attacks for image classifiers as well as for semantic segmentation. Furthermore, we proposed methods for adversarially training neural networks to become robust against universal perturbations and universal adversarial patches. Moreover, we provided methods for test-time adaptation of neural networks to improve robustness to domain shifts and studied the role of shape-biased representations on robustness to common image corruptions.
A different strand of my research is automating machine learning (AutoML), specifically Neural Architecture Search. The latter research field is motivated by the vast design space of neural networks and the diversity of inference hardware. Manually tailoring a neural architecture for every type of hardware is cumbersome and not scalable - hardware-aware neural architecture search can vastly improve design efficiency and thus reduce cost of AI development. See our survey on Neural Architecture Search and a more recent survey on Neural Architecture Search for Dense Prediction Tasks in Computer Vision. We have also developed AutoCLIP, a method auto-tuning zero-shot classifiers for vision-language models that improves zero-shot performance across a broad range of domains.
I also love contributing to machine learning libraries, both open source and proprietary. I am a (recently mostly inactive) core contributor of scikit-learn, where I contributed tools for probability calibration of classifiers and for kernel ridge regression. Moreover, I have written a complete redesign of the Gaussian process module for scikit-learn. At BCAI, I was involved as a core developer of frameworks for deep learning training pipelines, neural architecture search, and robustness evaluation.
I am an ELLIS member and an ELIZA Industrial Fellow. I serve as an Area Chair for NeurIPS 2026 and regularly review for scientific conferences and journals such as ICLR, ICML, NeurIPS, and TMLR. I was senior area chair of the AutoML 2022 conference and co-organizer of the workshops NAS@ICLR 2020 and NAS@ICLR 2021. I have been recognized by ICLR as Highlighted Reviewer in 2022 and Outstanding Reviewer in 2021.
See my personal website for more details.