Microservices are exceptionally painful when building machine learning models on complex data sources like user behavior. In these cases, to make a prediction, you need to pull from a multitude of microservices, that in turn pull from another multitude of services to get all the contextual ...