If you are already doing Declarative Pipelines you can indeed encapsulate the transformation logic a view inside the pipeline. That view would be reading from a source (lets say) silver streaming table and populating a (lets say) gold MV.

So raw files (autoloader) --> bronze (append streaming table) --> silver (SCD streaming table) --> view (some additional transformations, joins) --> gold (MV).

This is already possible.

Maybe I'm misunderstanding the request. Feel free to clarify.