We have used machine learning to study stellar rotation in TESS despite the mission's complicated systematics. The upcoming Nancy Grace Roman Space Telescope will perform time domain surveys at multiple wavelengths that stand to increase the number of period measurements and offer temperature resolution for star spot properties, shedding light on the connections between rotation and magnetism. However, the survey design is not yet decided, and certain choices may be critical to ensure sufficient cadence, baseline, and wavelength coverage for stellar rotation science. We are using the simulation and machine learning framework developed for TESS to predict the optimal Roman survey design for stellar rotation. I will discuss our framework and illustrate how existing machine learning tools can inform decisions for survey design. I will consider the stellar populations and periods Roman will be sensitive to and preview the transformational science Roman will enable. Theme(s): The Sun and Cool Stars in the Time Domain

Also presented 2024 July at the TASC 8 conference in Porto, Portugal.