As the power system moves from thermal plants (synchronous generators) to wind and solar (inverter-based), the dynamics of the grid become increasingly dependent on the dynamics of power electronic-based inverters. The fundamental challenge preventing the inclusion of detailed inverter models in interconnection scale simulations is the computational burden of solving such a large differential algebraic system. This research project aims to leverage recent advances in scientific machine learning to speed up the solution of large, stiff dynamic systems. In the example shown for the single machine-infinite bus system, the differential equations which define the system are replaced with a neural network which is trained to learn the dynamic response of each state variable. The end goal of this project is to demonstrate novel scientific machine learning techniques on a much larger scale, including verified models of the Maui power system.