M.Sc in Mechanical Engineering Department, Columbia University
Panasonic AI Lab, Panasonic R&D Company of America
He is mainly conducting research on robotics learning using reinforcement learning. Especially, he focuses on the "sim-to-real" and the "image-based RL" aspects of robotics which have a huge potential to release the human from the intense labor in the manufacturing field.
He has developed DoorGym, the RL agent training environment focused on a door-opening task, with supervision from UC Berkeley and it has been open-sourced online. DoorGym was presented in the NeurIPS 2019 Deep Reinforcement Learning workshop.
His interests lie within actuating and control the physical object. He enjoyed developing the autonomous control system of the MagLev train when he used to work for a railway company. Even he switches his career from train to robotics, his passion still continues. The reason he likes robotics is that it is a multi-disciplinary field that he has to wear a lot of hats to make the thing works.