Dynamic Digital Tumors for Precision Oncology is a collaborative project focused on transforming cancer care through next-generation AI technologies. This initiative aims to develop advanced Drug Recommender Engines (DREs) that match patients with breast, lung, or colon cancer to the most effective therapies. One key innovation is the Genotype-to-Phenotype Transformer (G2PT), which models complex relationships between genetic variants and disease traits to enable fine-grained prediction and interpretation of drug response. By integrating data from genomics, health records, and medical imaging, these models will predict therapy outcomes, explain drug resistance, and adapt dynamically as new information becomes available.