A Gaussian process (GP) is a powerful Bayesian statistical learning method that allows complex, non-linear functions to be modeled probabilistically. Rather than try to guess at an appropriate parametric model, this approach models the underlying function directly, using Gaussian (normal) distributions. How can something as simple as a Gaussian distribution be used to model something so complex? Come along and find out! I will demonstrate how to use GPs for both regression and classification tasks. I will introduce the Gaussian process for new users, as well as the PyMC3 package for Bayesian modeling. We will then use the new Gaussian process module for PyMC3 to solve a variety of statistical and machine learning problems.