Hmm... I assumed it was google-backed because the author has published a few of the algorithms in well-known conferences using his google email, but yes, this code is probably not used extensively, if at all, internally.