Cell-based therapies are a frontier of regenerative medicine, offering the potential to regenerate damaged or diseased tissues to a fully functional status.  Stem cells are a particularly promising source of cells for therapeutic applications because of their high self-renewal potential and ability to differentiate to multiple cell types.  However, for stem cell-derived cellular therapies to meet their promise, robust and scalable processes are needed to differentiate the stem cells to high quality somatic cell types.  My lab focuses on developing simple and efficient processes to guide human pluripotent stem cells (hPSCs) to a variety of fates and implementing scalable platforms to generate cells at the quality and number needed for therapeutic applications. In this talk I will focus on our efforts in developing defined methods to direct hPSCs to cardiovascular cell types, including cardiomyocytes, endothelial cells, epicardial cells, and cardiac fibroblasts.  I will also discuss efforts to transition from 2D lab-scale differentiation processes to 3D processes capable of generating clinical doses (~10⁹ cells).  We have found that generating tissue-like structures containing multiple different cell types (e.g. cardiac fibroblasts and cardiomyocytes) increases the functional properties of the cardiomyocytes and have developed methods to integrate manufacturing of 3D aggregates containing multiple cell types in suspension.  Finally, I will discuss efforts toward developing closed-loop cardiomyocyte manufacturing processes.  Through a multi-omic analysis of hPSCs differentiating to cardiomyocytes we have identified gene, protein, and metabolite features that allow us to predict the purity of cardiomyocytes at the end of a 15 day differentiation batch as early as days 2 and 4, enabling identification of failed batches early in the process and perhaps facilitating rescue of these batches through closed-loop interventions.