At some point, people realized that 90% of data science work was building pipes and keeping them clean. And then data engineering was invented so that the data scientists wouldn't have to get their hands as dirty. :)