We have different use cases. Our banking use case uses machine learning to identify customer life events and recommend the best-suited card products. These machine-learning models are deployed in our environment, where they run on a scheduled basis. We rely on the platform for every data science modeling. We have 170 models in production. It is very helpful for production and evaluation. 

I am a professor and this is one of the solutions that I use as a teaching tool for my students.

The most recent version can be used by the students while they are working in the labs because our hardware has sufficient RAM to run it. However, many of the students have laptops that do not have enough RAM, so they cannot work with it outside of the labs. In this case, the students are able to use an older version with lesser demand on hardware requirements.