Jamie Meindl walks through PyTorch, the open-source deep learning framework, as a supplementary session for MIT's 6.7960 Deep Learning course. The tutorial assumes general programming background but no prior PyTorch experience, and covers the practical mechanics of building and training neural networks in the framework: tensors, autograd, model definition, and the training loop that students will use in the course's problem sets. Delivered in a code-along style typical of MIT OpenCourseWare supplementary materials rather than a formal lecture hall talk, it is meant to get students functional in the tool quickly rather than to argue a thesis. Useful as an onramp for anyone who knows the underlying deep learning theory but has not yet written PyTorch code.