I am a solution architect and a consultant, and I use H2O as a machine learning platform. I create ensemble models using R and H2O, tune the hyperparameters, and then deploy them.
There are various use cases for this solution. One of the ones I worked on was a trailer forecasting solution. The customer wanted to understand the preload capacity that would be needed to have on hand so that they could call upon the right sized trailers and the right packages. It was a problem of logistics where you had to determine how many trailers were required in order to ship the packages being transported, and also have them ready just in time.