I have 3 postdoctoral openings for which the official ad will come out soon. Start date: As soon as possible, ideally not later than January 2021. Duration: Minimum one year, extendible to two years depending on the outcomes of the first year.

The post-doctoral appointee(s) will conduct research on theoretical aspect of efficient and effective learning and planning in sequential decision making under uncertainty. The positions are supported through NSERC and a Canada CIFAR AI Chair grant and the candidate will work with Csaba Szepesvari and his research group. The work involves original theoretical research, collaboration with and supervision of PhD and MSc students. Specific topics include (but are not limited to):

  • computationally efficient and effective online learning and planning in large MDPs, or with batch data

  • limits of such algorithms

  • new algorithms for multicriteria reinforcement learning

  • efficient optimization algorithms for reinforcement learning

  • policy performance certificates.

Candidates must have demonstrated excellence in learning theoretic approaches to reinforcement learning or closely related learning problems through publications at relevant venues such as COLT, NeurIPS, ICML and alike. Background in either mathematics, computer science, optimization, statistics, operation research, or control is required.

Current group

PhD Students

MSc Students


Post-doctoral fellows, research associates

PhD students

  • Zoltán Szamonek ELTE, 2007. Zoltan decided to quit his program and since then he is an SE at Google in Zürich.

MSc Students

  • Shun Jie Lau (2012-2018). Present position: Software Developer at CEMWorks Inc.

  • Zoltán Kömives ELTE; co-supervised with András György 2007.
    Position after graduation: IT Consultant at SK Point. Currently Software Engineer at J.P. Morgan & Chase, UK, Glasgow.