I am a research scientist at the Systems Technology Lab of Adobe Research. My main research interests are in statistical machine learning from volatile Web data, especially, from user-generated content and user behavior.

Most recently, I am working on scalable reinforcement learning algorithms for sequential decision making in digital marketing. Employing Spark, Hadoop, and related technologies belongs to the day-to-day engineering processes in order to handle the huge amount of data describing customers and advertising campaigns. In our team, we combine these state of the art technologies with insights from our research in order to automatically infer advertisement strategies in a timely manner that go beyond current targeting models.

Prior to joining Adobe in November 2012, I worked as a research associate at the Web Technology and Information Systems group at the Bauhaus-Universität Weimar, Germany, where I developed expertise in data mining, machine learning, and text classification. In 2011, I was hosted by the machine learning consulting group "The Church and Duncan Group, Inc." in San Francisco and conducted research on new algorithms for look-alike modeling in the digital advertisement context. I published at several international conferences (e.g., ICDM, WWW, CIKM, SIGIR, ECIR, CLEF) and served frequently in the program committees of TIR, PAN, and I-KNOW. Most of the publications can be found here.