How did you calculate that ROI? To me it sounds like an excellent investment. The company invested 10% of your time, and gained 100% extra time (assuming the new hire's time is worth as much as your own).

It is the sort of ROI that can scale a company from 20 to 20000 employees in just a couple of years :)

I think that's the wrong way to look at it. The time she put in also meant that some people that could potentially have been a bad fit were not hired which was good.

While Google goes straight from application to a telephone interview, a lot of companies ask for a work sample test first - either by e-mailing them a problem and asking them to reply with a solution, or through automated systems like Codility.

If you're spending a lot of employee hours filtering candidates who could have been filtered more easily through other means, I can understand feeling a bit indifferent about it.

True, but based on her little "where are they now" aside, she could have just randomly selected 98%* of the candidates she interviewed for rejection and achieved nearly the same end result.

(or whatever the actual percentage was, I don't know the total)

Exactly, and maybe the good ones did get hired and really knocked it out.

That said it sounds like the quality of candidates sent to her were really low and the upstream recruiters should have been working to improver that.

Note, however, that this may not be what the good engineer wanted to be doing. (Also, just not putting bad candidates on the good engineer's plate would be more efficient.)

Depends on how good that other person is, doesn't it?