Pie & AI: Madrid – Playing with reinforcement learning