Bruno Scherrer has worked on reinforcement learning since 1999, contributing to the theory of approximate dynamic programming algorithms and complexity of the Policy Iteration algorithm.