As usage of the Jupyter notebook in scientific computing becomes increasingly ubiquitous, so too does its presence in classrooms. In this talk, I will describe nbgrader, a tool developed by and for instructors to create and grade rich, interactive assignments in the notebook. On its own, nbgrader provides functionality for creating assignments and then for both automatic and manual grading of submissions. When combined with JupyterHub, it supports the full grading pipeline: creating assignments, releasing them to students, collecting submissions, grading, and generating personalized feedback. To demonstrate the use of nbgrader, I will walk the audience through the basics of the tool, and then outline two example workflows: one using standalone nbgrader, and one using nbgrader with JupyterHub. In doing so, I will illustrate to instructors how they can create their own assignments in the notebook using nbgrader.