PhD
Chair of Data Analytics and Machine Learning at TUM
Informatik 26
Department of Computer Science
Boltzmannstr. 3
85748 Garching
Biosketch
Tim Tomov is a Ph.D. student in the Data Analytics and Machine Learning Group at TUM under the supervision of Prof. Stephan Günnemann. He received his Master’s degree in Mathematics in Data Science from TUM, where his thesis focused on uncertainty quantification under ambiguity. During his Master’s studies, he completed a Research Scientist internship at Amazon and conducted research on scene synthesis in Prof. Matthias Nießner’s Visual Computing Group. Prior to this, he was a Research Assistant in the Image-Based Biomedical Modeling Group under Prof. Bjoern Menze, where he worked on medical deep learning. He also gained professional experience in industry through roles including Product Owner and Associate Data Scientist.
relAI Research
Uncertainty Quantification for Reliable Decision-Making
Reliable decision-making with AI systems requires uncertainty estimates that are meaningful for the decision at hand. Tim’s research focuses on uncertainty quantification as a foundation for reliable decision-making, studying when models are uncertain, what kind of uncertainty they express, and how this uncertainty should inform downstream decisions.
A central focus of his work is uncertainty quantification under ambiguity, where multiple outcomes may be valid, and uncertainty cannot be interpreted only as a lack of knowledge. He also studies uncertainty quantification in structured tasks, where predictions consist of multiple interdependent components and uncertainty must be assessed with respect to the overall task. His current work investigates these questions in the context of large language models. The goal is to develop uncertainty-aware methods that make AI systems more reliable in complex, decision-oriented settings.