he paradoxical nature of T-cell receptor (TCR) specificity, which requires bothprecise recognition and adequate coverage of antigenic peptide-MHCs (pMHCs),poses a fundamental challenge in immunology. Efforts at modeling this complexmany-to-many mapping have focused on the detection of reactive TCR-pMHCpairs as a binary classification task, with little success on unseen epitopes. Here,we present TCR-TRANSLATE, a framework that adapts low-resource machinetranslation techniques including semi-synthetic data augmentation and multi-taskobjectives to generate target-conditioned CDR3b sequences for unseen inputpMHCs. We examine twelve model variants derived from the BART and T5model architectures on a target-rich validation set of well-studied antigens, find-ing an optimal model, TCRT5, that samples experimentally validated CDR3b se-quences for unseen epitopes. Our findings highlight both the potential and lim-itations of sequence-to-sequence modeling in rapidly generating antigen-specificTCR repertoires, emphasizing the need for experimental validation to bridge thegaps between predictions, metrics, and functional capacity.