My research interests lie at the intersection of Statistics (time series analysis, non-parametric regression) and Computer Science (machine learning and data mining). I am particularly interested in interdisciplinary work and applications of computational methods to real-world data, e.g., from electrophysiology, distributed physical systems and economy. In my PhD work at the University of Luebeck, Germany, I studied properties of order statistics in long-memory processes with applications to segmentation and clustering of electroencephalography (EEG) recordings. After my PhD, I spent two years as a Postdoctoral Research Fellow at the University of Waterloo, Canada, where I was working on algorithms for predicting human activities from low-level sensor measurements, and theoretical foundations of Conditional Markov Chains. In 2011, I joined IBM as a Research Staff Member where I have been working on scalable systems for real-time predictions of transportation and energy systems. Since 2013 I am leading the Exploratory Predictive Analytics team in the IBM Dublin Research Lab, focussing on the development of flexible, robust, high-dimensional regression methods and their application to real-world data.
I am the co-author of more than 30 peer-reviewed publications and co-inventor of four US patents. I was an invited keynote speaker at the 2014 ACM e-Energy conference. My research has been founded by the German Research Foundation (DFG), the Canadian Bureau of International Education, Google Inc and MITACS. I have been a reviewer and program committee member for numerous journals and conferences in data mining, machine learning and artificial intelligence, including IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Transactions on Smart Grids, Data Mining and Knowledge Discovery, ICML (2013-2014), IJCAI (2011, 2013), NIPS (2010, 2013-2017), UAI (2011, 2014-2017).