Scientists led by Hiroyuki Hamada at Kyushu University have leveraged ProtGPT2, a protein language model, to design smaller, novel variants of malate dehydrogenase enzymes for bioprocessing. This AI-driven approach identifies functional motifs critical for enzyme activity and proposes structure-stabilized enzyme candidates, showcasing strong similarity to natural proteins. Early structural analyses confirm active site preservation, suggesting these AI-generated enzymes could optimize bioprocess workflows by improving functional enzyme design and performance.