Data Science has enabled companies to establish predictive models about their sales, forecast their needs in human resources, enhance theirjavascript:void(0); customer knowledge and so much more. But what is the afterlife of these models ? Are they doomed to perform one-shot predictions and then fade away? After the training and testing steps, the final part of an end-to-end Data Science project should be deploying the constructed model in a production environment in order to reuse easily its results.

Today this step can be quite time-consuming when it involves rewriting completely the machine learning model in another language or combining specific skills in machine learning and production coding. In this talk, we will present several techniques to automate the deployment of any R model in two complementary production environments : a big Data cluster and a web service.