Researchers led by Nobel Laureate David Baker have developed AI-driven design of highly specific protein binders that mimic T cell receptors targeting peptide-MHC complexes on diseased cells. Published in Science, the approach creates low-immunogenicity binders targeting diverse viral and tumor peptides, including challenging antigens like PRAME. This technology accelerates scalable personalized immunotherapy development, potentially overcoming the allele diversity limitations of natural TCRs. The work leverages diffusion models and structural validation, marking a major step in AI-enabled de novo protein design for oncology applications.