System is designed to use the Jetson TX2 platform as a sensor fusion center. Computer vision inference and RFID work collaboratively to train and improve models. Object classes are added as RFID tagged products are introduced into the store. Products are associated with object class data via RFID tags attached at initial phases of production. This approach allows for the system to phase in products in a AmazonGo-style retail environment. Over time the RFID tagging can be omitted once the CV inference performance is above the threshold approved by the merchant.