The light traveling from a distant object can be bent by the mass of an object between the two: this effect is called gravitational lensing. Upcoming astronomical surveys will capture an unprecedented number of lensing systems in the Universe. While deep learning is showing promise in detecting and quantifying these distortions, robust modeling of synthetic lensing images is needed. A team at Argonne is training a modular deep learning pipeline for denoising and deblending these lensing systems.