Sara Beery lectures on transfer learning as part of MIT's 6.7960 Deep Learning course, Fall 2024. She covers how models trained on one data distribution can be adapted to another, focusing on generative models used as data augmentation tools, domain adaptation techniques for bridging gaps between source and target datasets, and prompting methods that let large pretrained models be steered toward new tasks without full retraining. The lecture builds on earlier course material on neural network architectures and training, applying those foundations to the practical problem of limited or mismatched data. Runtime is 76 minutes, consistent with a full graduate-level class session, and the material assumes familiarity with deep learning fundamentals covered earlier in the course.