| Abstract: Active galactic nuclei (AGN) accretion disks fuel powerful AGN feedback and are important laboratories for studying accretion physics. Because AGN are generally too distant to resolve spatially, most of what we can learn about their accretion disks comes from studying the variability in AGN light curves. However, many mysteries remain about the source of variability in AGN light curves and what this variability can reveal about the structure, internal physics, and accretion rates of AGN disks. I will show how I use radiation magnetohydrodynamic simulations, long baseline AGN light curves, and machine learning tools to improve our understanding of AGN variability and model the structure of AGN disks. These models are crucial for understanding AGN at all redshifts, including the high redshift little red dots recently discovered by the James Webb Space Telescope, which may be super-Eddington AGN. Link to the Event Video |