Superb Explanation @Vincent .

The ML deployment is really a subjective and use case specific. You have explored practical approaches and their directions which are quite interesting.

Thanks for wonderful post 😊