Deployment of trained model on to cloud side requires to support clients with varying languages. REST API appears to be one such a method in which Client can communicate with Inference Engine which is performing inference for a given input data. Where, input data come from Client applications which are written in different language

tfServingDockerRestV2