I can definitely see it speeding up manual efforts along the data processing pipeline, but data engineering is more than schema generation, just as NLP is more than feature extraction from structured documents.

Very interested in what new tools and workflows would emerge from this though!