Dielectric optical coatings are crucial for ultrafast laser systems. Current methods for designing these coatings rely on brute force optimization, which is computationally intensive and demands expert knowledge. Our project introduces an Al-driven approach using deep neural networks to solve this complex inverse problem, offering faster, expert-independent solutions. Promising first results have been obtained for simple coating designs.