The story starts off saying that the work's being done by analysts:
> The work involves having analysts trawl through social media sites to look for what the document describes as “inflammatory” postings and then sharing that information across government agencies.
, so it sounds more like they've got some folks browsing social-media.
Granted, a lot of the post-scanning would seem better done by bots, and stuff like sentiment-analysis could help classify posts for human inspection, so they'd probably want to hire a few engineers, but why hundreds? And why 250-500 kUSD/yr for such mundane work?