I’m a sort of machine learning scientist applying and explaining classical ML models on tabular data to generate insights, and have done lots of research on dis

I’m a sort of machine learning scientist applying and explaining classical ML models on tabular data to generate insights, and have done lots of research on disciplines in data science. I think by now, we should recognize that data science is just the broad field of extracting insights from data, encompassing - traditional statistics - all kinds of exploratory and other data analytics - machine learning ... where the primary goal is, again, extracting insights from data. On the other hand, if you are building something, eg products or data pipelines, you are less of a (data) scientist and more of an engineer. Sometimes this may overlap, eg, you train some models to get insights (data science) but then you also productionize and deploy it (ML engineer). In bigger orgs, these should probably be separate teams. Regarding how common the switch is, I dunno. Personally, I think that increasingly the more serious teams and companies are recognizing that data science is a bubble (often just renamed analytics) and what has been, is, and will continue to be more important, is engineering. Eg data engineering to tame your data and ML engineering to build and deploy products. Data scientists are in the middle or on the side to develop POCs, run experiments, and help action any insights, in close collaboration with product and business. So that’s my take on this.