I highly doubt machine learning is going to become a commodity skill any time soon. The optimal model structure is tightly coupled to the specific data and the types of questions you want to answer. Additionally, understanding your model well enough to really trust it requires significant knowledge.

I expect that common deep learning tasks are going to end up being performed by large pre-trained networks behind an API. Image tagging, sentiment analysis, etc. It will be possible to fine-tune these networks for specific tasks by retraining top layers using new data.