You manage a team of data scientists who frequently submit training jobs to a cloud-based backend system. Over time, managing this system has become challenging and cumbersome. You are considering switching to a managed service to simplify administration. The team utilizes a variety of machine learning frameworks, including Keras, PyTorch, Theano, Scikit-learn, and some custom libraries. Given these requirements and the need for a more manageable solution, what should you do?