Ksenija Stepanovic

Greetings, fellow internet voyager! I am Ksenija. At TU Delft, I am part of the Flex-Heat project alongside companies Withthegrid and Vattenfall. Our quest is to use the powers of artificial intelligence to forge a path toward sustainable energy transition. However, that path is challenging. The control of energy systems requires algorithms that are not just efficient but also safe, fast, and scalable.

Over the past three and a half years, my focus has been on the research and development of data-driven controllers based on the above requirements. These controllers have been developed through my research on the integration of artificial intelligence with physics-based models. More specifically, they involve advancements in compiling learned deep neural network models in mathematical optimization, constrained gradient descent optimization, and relaxation of deep neural network models. The practical impact of my work shines through in the realm of district heating systems, where these controllers have reduced operational costs and CO2 emissions and increased the integration of renewable energy sources. Beyond the specific domain of district heating, these controllers hold potential for optimizing the operations of any energy system with complex state transition functions.

In addition to designing and developing algorithms and programming, I love traveling, hiking, and spending time in nature. My passion extends to exploring new cultures and immersing myself in diverse environments. Moreover, I am deeply interested in the art of communication, team dynamics, and negotiations. My firm belief is that an inclusive, open and kind environment is the cornerstone of any success. That is why I am also a Confidential Officer within the Algorithmics group.

If you are working in any adjacent field and are interested in having a talk, please don't hesitate to contact me at K.Stepanovic@tudelft.nl.