DP is most definitely not in fashion because of reinforcement learning. It's been pretty standard since long before the ML buzz.