Massive astronomical telescopes are utilized by observatories and institutions all around the world to make exciting discoveries. During operation, the coordinates of a desired celestial object are programmed into the telescope; which ideally, then tells the telescope mount to move and center its image-collecting area on that exact spot in the sky. However, due to a variety of factors like atmospheric conditions or offsets in the motors, this pointing maneuver is not an exact function. Without any corrections, the image produced by the telescope will be off-center, or potentially not in the frame at all. None of the current methods to correct for this error involve any sort of artificial intelligence component, and most rely on a slow process of human intervention to manually adjust the instrumentation. Running these machines costs thousands of dollars a night, and time spent re-aligning the telescope through slow, manual adjustments is costly. In this talk, I will talk about our development of a novel pointing system, built with an underlying recurrent neural-network architecture. I will also discuss the inherent challenges in training a model of this nature, and how we can adapt the recurrent model to a variety of scenarios. We have already seen success working with the WIYN telescope at Kitt Peak Observatory, and we believe that this modern approach can be fit to potentially hundreds of telescopes around the world.