As an example, we describe a use case applying geophysical simulations and inversions for subsurface imaging. This involved running 1000 1D inversions that each produced a layered model of the subsurface conductivity, which were then stitched together to create a 3D model (to understand why the Murray River in Australia was becoming more saline). This combined simulations, data analysis, and machine learning with interactive visualization. The initial workflow was developed on the user's  laptop environment, and needed to be scaled up at NERSC.