This document discusses model serving for deep learning. It begins with a brief introduction to machine learning, deep learning, and neural networks. It then explains that deep learning has a growing impact and can perform better than other machine learning techniques and humans. The document focuses on model serving, including what a deployed model looks like, key aspects of model serving systems like performance, availability and monitoring, and examples of model serving systems. It describes the Amazon Model Server and its features like model archives, REST APIs, containerization, metrics, and ONNX support. In closing, it discusses challenges and opportunities in model serving.