To address challenges of fresh waste and spoilage, we are proposing to develop AI based predictive solutions based on following sub-aims: (a) AI enabled rapid detection and quantification of spoilage microbes of fresh produce in a limited resource environment; and (b) AI enabled prediction of spoilage risks based on ripening status, presence and quantity of distinct spoilage microbes and the expected transport and distribution of the selected food products and (c) semi-supervised domain generalization for the above aims. The efforts in Aims 1 and 2 are integrated with Aim 3 to enable domain generalization and data efficient AI models.