My research aims to develop adaptive controllers that leverage real-world data for online improvement of dynamic models for soft robotic systems during policy learning to reduce the gap between simulation and real-world performance.