What fields would you guide them to instead?

I may be one of the developers you speak of (with academic aspirations), presently considering my path forward.

I'm sceptical if going down the ML swamp is the best way forward.