The lack of realistic in vitro organ models that can faithfully represent in vivo physiological processes is a major obstacle affecting the biological and medical sciences. The current gold standard is animal experiments, but it is increasingly clear that these models mostly fail to recapitulate the human physiology. Moreover, animal experiments are controversial, and it is a common goal in the scientific community to minimize the use of animals to a strictly necessary minimum. The emergence of stem cell engineered organ models called organoids represents the only viable alternative to animal research. However, current organoid technology is yet to produce the larger physiologically relevant organmodels that the medical sciences really need. Specifically, current organoids are too small, not vascularized and lack the 3-dimensional organization found in vivo. In this interdisciplinary project we aim to challenge all these limitations by using the recently developed gastruloid technology guided by cutting edge bioengineering and artificial intelligence. Gastruloids are formed by initiating the very early developmental processes and develops along a highly coordinated three axial process that closely resembles mammalian embryogenesis. Moreover, gastruloids can develop several organ precursors simultaneously and thus constitutes important improvements over conventional single-tissue organoids. To harvest the potential of gastruloid technology we will first implement large sequencing and imaging experiments to optimize the developmental trajectory of gastruloids for organ inductions. We will then build these datasets into a multimodal data matrix to identify gastruloid candidates for cardiovascular and foregut development. Specifically, we will identify candidates that show strong vasculogenesis as candidates for later vascularisation by anastomose with endothelial cells.