The design and analysis of symmetric cryptographic primitives requires often solving combinatorial optimization problems over very large search spaces, that usually cannot be explored exhaustively. In this lecture, we show how some of these problems can be solved using metaheuristic techniques, namely Evolutionary Algorithms (EA). Specifically, we show the use case of automatically designing Boolean functions and S-boxes for stream ciphers and block ciphers, and how to apply Grammatical Evolution to construct neural networks for Deep Learning-based Side-Channel Analysis.