Applied biocatalysis is driven by environmental and economic incentives for using enzymes in the synthesis of various pharmaceutical and industrially important chemicals. Protein engineering is used to tailor the properties of enzymes to catalyze desired chemical transformations, and some engineered enzymes now outperform the best chemocatalytic alternatives by orders of magnitude. Unfortunately, custom engineering of a robust biocatalyst is still a time-consuming process, but an understanding of how enzyme function depends on amino acid sequence will speed up the process. This review demonstrates how recent advances in ultrahigh-throughput screening, mutational scanning, DNA synthesis, metagenomics, and machine learning will soon make it possible to model, predict, and manipulate the relationship between protein sequence and function, accelerating the tailor design of novel biocatalysts.