Botond Cseke

Research Scientist

Botond has completed a PhD in Machine Learning at Radboud University Nijmegen under the supervision of Tom Heskes where he learnt not care about anything that cannot be described by a probabilistic graphical model. This was followed by two post-docs at the University of Edinburgh (bio-informatics) and Microsoft Research Cambridge (approximate inference), respectively. He joined the Volkswagen AI Research lab in October 2017 where he enjoys modelling with probabilities and experimenting with inference in probabilistic graphical models using neural networks or deep neural models, whatever you fancy to call it. In his free time he does another type of modelling


  • variational inference in graphical models
  • variational inference in time series models
  • generative adversarial networks