Auto Train Hierarchal Labeled Data Set

What is the best way to take advantage of data that has been labeled with nested hierarchies?

For example I will have a sentence:
“The patient named Ricky and is coming in for a cough.”

here the high level span is called [name span] which captures the span where the name is used so in this case the [name span] is: {patient named Ricky}.

The name span is the top level hierarchy though, so there are two labels in the lower level hierarchy: [first name] and [last name]. They respectively capture the first name and last name. So in the example sentence, there is no last name, just a first name.

So the word {Ricky} would have the [first name] label applied.

Can I use AutoTrain to take advantage of the data already being labeled this way? I know I can break them out as independent models, but I was hoping their could be a more advanced method that allows for the model to take advantage of this hierarchical information