This tutorial will cover practical approaches to some of the most pressing challenges in medical imaging today. We’ll look at how to work with limited labeled data—using classic tools for label generation, one- or few-shot learning, and training with synthetic data. We’ll also talk about how to handle domain shifts between datasets, including the use of data augmentation and domain adaptation methods. Finally, we’ll introduce implicit neural representations and show how they’re being used for more flexible and accurate image reconstruction.