Research Interests
Symmetry-Informed Machine Learning, Information Geometry, Differentiable Programming
Professional Affiliations
- DNTOX, Co-founder Affiliate
- IUF, Research Affiliate
- RUB, Research Affiliate (alumnus)
- BUW, Research Affiliate (alumnus)
Social Profiles
Contact Information
I am a mathematician and informatician developing foundational mathematical frameworks to build more robust, efficient, and explainable artificial intelligence.
My primary research program is focused on solving complex problems where standard "black-box" AI models are insufficient. I develop and apply methods from symmetry-informed learning, information geometry, and differentiable programming to create systems that can reason about underlying structure. This allows them to generalize effectively from sparse or noisy data and deliver explainable results—capabilities that are critical for high-stakes applications in scientific modeling, hazard assessment, and cognitive science.
I put these foundational principles into practice to solve critical real-world problems. As a co-founder of DNTOX, I apply these approches to advance environmental health. We are on a mission to replace animal testing by AI-driven methods to screen chemicals for neurotoxicity.
My academic background includes a PhD in mathematics from the Ruhr-University Bochum, where I applied geometric-invariant theory to classification problems, and postdoctoral research at the Leibniz Research Institute for Environmental Medicine (IUF) and the Ruhr-University Bochum (RUB).
I am always open to discussing shared interests and potential collaborations.
Working Papers
A comparative study of biostatistical pipelines for benchmark concentration modeling of in vitro screening assaysNew Paper published in Computational Toxicology
Abstract. New approach methods (NAMs) have been prioritized to reduce the use of animals for chemical safety assessment while continuing to protect human health and the environment. A key challenge of generating toxicity data is the implementation of a standardized analysis approach for transparent and reproducible benchmark concentration (BMC) estimation and uncertainty quantification for assay developers, regulators, and other stakeholders. In this study, we compared the bioactivity results of 321 chemical samples from four established BMC analysis pipelines used for evaluation of developmental neurotoxicity (DNT) NAMs data: the ToxCast pipeline (tcpl), CRStats, DNT DIVER (Curvep and Hill pipelines). We found an overall activity hit call concordance of 77.2 % and highly correlated BMC estimations (r = 0.92 ± 0.02 SD), demonstrating generally good agreement across pipelines. Discordance appeared to be explained predominantly by noise within the data and borderline activity (activity occuring near the benchmark response level). Evaluation of the BMC confidence intervals indicated that pipeline selection may impact the estimation of the BMC lower bound. Consideration of biphasic models appeared important for capturing biologically-relevant changes in activity in the DNT battery. Lastly, different approaches to compute ‘selective’ bioactivity (activity below the threshold of cytotoxicity) were compared, identifying the CRstats classification model as more stringent for classifying selective activity. Overall, these findings indicated greater confidence in NAMs bioactivity results and emphasize the importance of understanding strengths and uncertainties of concentration–response modeling pipelines for informing biological interpretation and application decision making.