This presentation delves into the application of scientific machine learning (SciML) to meet the increasing computational demands associated with computational physics. As simulations become increasingly complex, there is a pressing need for high-performant and fast code to efficiently handle computational workflows. Through an in-depth examination of SciML techniques, we will discuss how the utility of scientific machine learning can lead to a significant improvement in the speed and accuracy of simulations. Additionally, this presentation will address the potential challenges associated with implementing scientific machine learning in computational physics workflows. From data availability, model interpretability, to uncertainty quantification, we will explore the potential challenges associated with SciML simulations and practical strategies to overcome them.