Those full stack roles are still relatively common (though decreasingly so, as you pointed out) at companies with small data science orgs, on the order of 10 pe

Those full stack roles are still relatively common (though decreasingly so, as you pointed out) at companies with small data science orgs, on the order of 10 people. But they're still usually called "data scientist," while I think of more specialized roles like MLE, ML ops, data engineer, data ops, etc. as cropping up in bigger and/or more mature data science teams. Calling it MLE instead of data scientist could be a screening technique though.