In this class, we look at evolutionary methods for the optimization of policies represented as Artificial Neural Networks. We cover direct and indirect encodings, and implement an algorithm based on the neural networks package Flux to learn control policies for the Gym CartPole task.