Real World Robot Learning: Learned Rewards, Offline Datasets And Skill Re-Use