About Me


I am a third-year PhD Student at KTH Royal Institute of Technology where I am jointly supervised by Prof. Kevin Smith and Prof. Zeynep Akata (University of Tübingen). Previously, I graduated from ETH Zurich (bachelor’s and master’s degree), during this time I also had a one-year stint at Disney Research Zurich. After graduating, I embraced a two-year associate position at SAP SE (Deep Learning Centre of Excellence) where I focused on tailoring neural networks to the company’s needs.

Research Interests

I am mainly interested in computer vision and machine learning in general. During my master’s thesis at the Data Analytics Lab (supervised by Prof. Thomas Hofmann & Prof. Carsten Eickhoff, ETH Zurich), I became increasingly interested in medical applications.

In my PhD I am mainly concerned with medical computer vision. To me it is fascinating how MR and CT imagery can be processed with neural networks. A common problem that occurs is when some machine learning model is trained with data that is different from the data that is later used in a clinical setting, for instance, when a different device is used or the protocol is slightly different.

I am trying to contribute to the solution of this so-called domain adaptation problem. While there exists a bulk of research for still images, sequential modalities haven’t received so much attention from the research community. Accounting for the additional temporal resolution poses a substantial complication of the original task.

Social Media

  1. LinkedIn
  2. GitHub
  3. KTH


  1. Lennart Alexander Van der Goten, Tobias Hepp, Zeynep Akata, Kevin Smith: Conditional De-Identification of 3D Magnetic Resonance Images, BMVC'21 {arxiv}

Supervised Theses

Year Author Type Title Institution
2020 Bas Straathof MSc A Deep Learning Approach to Predicting the Length of Stay of Newborns in the Neonatal Intensive Care Unit KTH
2020 Oskar Danielsson MSc Multimodal Brain Age Estimation KTH
2018 Aziz Ben Othman MSc Enterprise-grade speech recognition service using open source software SAP SE
2018 Fabrizio Primerano BSc Design and Implementation of a Deep Learning Model to Automatically Colorize Black and White Videos Providing Temporal Consistency across Neighboring Frames. SAP SE