Great content! As a machine learning engineer I would say that there is no such a thing as the “best” method. It is important to be flexible and to adapt to each use case