This example should be the same as the merlin multi-stage recsys example. It should behave the same way. We should be able to take the ensemble created in that example and apply an offline runtime to create recommendations using the dask executor on a dataframe/dataset.