Hah yeah but this is just the reality of ML. "Engineer who only codes up awesome new ML algorithms" isn't a real job. For every 1 person coming up with exciting new algorithms, there's 500 people dealing with data pipelines, cleaning, maintenance, and operations. A while ago I got hired by a large SF tech company as a "machine learning engineer" and my team spent nearly all of their time writing Javascript web wrappers on top of scikitlearn. Thrilling stuff.
I've done this with mixed results as far as career satisfaction goes.
To be honest, getting pigeon-holed isn't always a conscious choice and once you realize that's what happened it's really hard to make a transition out of it. Sometimes the best approach is to play the cards you've been dealt.
As long as one does work which is personally meaningful, intellectually challenging, and which pays "enough" and has some stability, that all makes the sting of status-envy less of a bitter pill to swallow.
On the other hand there's no shortage of obsessively status-focused people out there. There's a guy on youtube whose channel is all about getting promoted at FAANG's, specifically, Amazon. That's it. Just practical deadpan advice on getting to "L7" (https://www.youtube.com/c/ALifeEngineered). I guess it's fine if that's what one _really_ wants out of life, but it seems like a recipe for profound meaninglessness for almost everybody.
This is unfair to DEs, not to mention inaccurate as of 2022. The statement assumes DEs are yearning to do "real ML" while suffering in their "data job" and waiting for a chance to move "upwards". Data engineering is a terminal career path by itself which may or may not support a dedicated ML function. In fact many of my DE colleagues switched from DS/ML/DL citing "model fatigue".
Now, "ML engineers" spending 90% of their time producing a workable dataset to get to the "real ML" is a different discussion.