From my point of view (I worked a few years with the language while in academia), there are no real job offers where Prolog makes a difference.
So maybe erlang people prefer a Prolog background but that's it (w.r.t. a job market perspective)
Given that Prolog isn't super-widely used (compared to say, C++), there are going to be a limited number of developers with Prolog expertise. That means companies who use it in their tech stack are generally going to have to hire developers who fit the other qualifications for the role then put them through some internal Prolog training. The listings for such positions probably won't have Prolog as a "required" skill but a "desired" skill.
One other point -- just because there aren't any/many jobs where Prolog is required doesn't mean there aren't many cases where Prolog could be useful. Prolog hasn't been widely taught for a couple of decades, so a significant proportion of developers today won't recognize a problem that Prolog would be a good solution for as a problem that Prolog would be a good solution for.
There's a Strange Loop 2014 talk on YouTube, "Production Prolog", that might give ideas about jobs that use (or could use) Prolog: https://www.youtube.com/watch?v=G_eYTctGZw8