Mission – Toward time series forecasting we can trust.
I currently focus on machine learning (ML) on time series, where the first question is usually: Is the model accurate? Plenty of research has since focused on constructing and selecting effective models, including some of my own. However, for many real-world applications, the more difficult follow-up is: Can we trust and depend on it? While substantial research has been conducted on classification, explaining, or even interpreting forecasting models remains largely unsolved.

I am opening up that black box to tackle these questions: What structure hides in the data in the first place? What did the model pick up? Can we use natural and formal languages as a bridge to the opaque domain of time series?

I am constantly drawn toward interdisciplinary collaborations to tackle real-world challenges using the vast toolbox of ML. In the past, I had the honor to work deeply with experts in the fields of finance, manufacturing, and oncology. You're more than welcome to reach out for new collaborations.

My research interests span other topics as well, including robotics, tractable probabilistic models (TPMs), graph neural networks (GNNs), large language models (LLMs), and causal discovery.

Application for Student Projects & Theses – Bachelor and Master.
I currently do not have the capacity to supervise additional theses or projects.

Timeline.

Supervised Theses (completed).

Teaching Assistant.