How did you finish this level of a reinforcement learning model in 48 hours?
The first few stages aren't terribly interesting, but the final stage where you swap roles after teaching your other "self" how to play is amazing. I think I managed 7 iterations before succumbing to my own creation.