I use Open AI's Gym and Google's Tensorflow to create a reinforcement learning AI that learns how to play Atari games.