A well-written article!

However, it presents google's solutions as the perfect ones, which is not necessarily right.

Even if Machine Learning as a research domain is old enough, the data scientist is still new as a role and profile. It's very normal that it will know evolutions in terms of tools, for example, and as you mention in your article, AutoML tools. the data scientists must obligatorily adapt their skills.

Besides this, the technical staff and solutions are just one part of a data scientist's task, his main task is to understand the client's businesses, and propose a model that adds value to its customers. In complex cases, this cannot be done by a machine or AI solution, this will always need a human brain.