In the High Performance Digital Geometry Processing Group, we develop novel natively parallel solutions to fundamental geometric problems at the intersection of computer graphics, scientific computing, and mathematics. Our efforts include devising new concurrent mathematical models as well as reformulating key serial models around vectorization from the ground-up when possible. To streamline the overall processing pipeline, we channel high performance on unstructured data through efficient linear algebra kernels. At the heart of this undertaking are new lean data structures for geometric representations, and efficient re-purposing of matrix algebra. In view of the rapidly growing field of machine learning, particular attention is paid to data-driven modeling.
For internships, HiWi positions, Bachelor and Master Theses: Please contact us for details.