The study presents a computational approach to design serine hydrolases, focusing on the catalytic triad and oxyanion hole mechanisms. Using RFdiffusion and ChemNet, the researchers generated proteins with increasingly complex catalytic sites and assessed their preorganization and activity. Experimental characterization revealed novel serine hydrolases with catalytic efficiencies up to \(3.8 \times 10^3\) M\(^{-1}\) s\(^{-1}\), close matches to design models, and distinct folds from natural enzymes. In silico selection based on active site preorganization across the reaction coordinate significantly improved success rates, enabling the identification of new catalysts from a small number of designs. The approach provides insights into geometric determinants of catalysis and offers a roadmap for designing industrially relevant serine hydrolases and complex enzymes that catalyze multi-step transformations.The study presents a computational approach to design serine hydrolases, focusing on the catalytic triad and oxyanion hole mechanisms. Using RFdiffusion and ChemNet, the researchers generated proteins with increasingly complex catalytic sites and assessed their preorganization and activity. Experimental characterization revealed novel serine hydrolases with catalytic efficiencies up to \(3.8 \times 10^3\) M\(^{-1}\) s\(^{-1}\), close matches to design models, and distinct folds from natural enzymes. In silico selection based on active site preorganization across the reaction coordinate significantly improved success rates, enabling the identification of new catalysts from a small number of designs. The approach provides insights into geometric determinants of catalysis and offers a roadmap for designing industrially relevant serine hydrolases and complex enzymes that catalyze multi-step transformations.