The scraping problems were medium hard and at least I learned some cool stuff along the way. All in all, it's been a pretty fun experience building the whole thing.
(ie.g. "www.agriculture.jobs" as example)
From my experience, lot of companies are willing to pay small amount of subscription if they have a "corp/niche specific" usergroup. (and this was actualy what my idea was about)
And if you are branch specific, it worth to curate the site, have some customer calls etc. - this would give you two winners: Companies looking clearly only for one set of applicants, and applicants who can be sure that the site content is valid.
Though, IIRC expeteer.com tried to deploy this "dating approach" onto job boards, but I do not know how they are doing these days