MIT chemical engineers unveiled an AI‑driven method that programs industrial yeast to boost production of therapeutic proteins, promising lower manufacturing costs and faster process development for biologics. The approach uses large language model techniques adapted to genomic sequence design to optimize expression and folding in yeast production strains. Parallel reports describe models that learn the ‘DNA language’ of yeast to predict regulatory sequences and increase protein yields. The technology targets key cost drivers in biologics manufacture by shortening strain engineering timelines and reducing experimental cycles. For biomanufacturing teams, the advancement signals growing utility of generative AI to accelerate strain engineering and process optimization; integration with existing fermentation and analytics pipelines could shorten timelines to clinical and commercial supply.