First DRL with Deep Q-learning