This is pretty interesting but I don't know what they were expecting. These hashing functions are pretty strongly rooted in mathematics, usually implementing the functions defined in math papers is the easiest part. Probably would have worked better if they got a bunch of math majors to do this instead of software developers.

I'm not trying to bash software developers either but most don't have the math background to write their own (good)hashing function.

Actually, for example the mathematical relationship between prime numbers and hash functions is unclear. Designing good hash functions seems to be an art relying largely on trial and error. This link has some details: http://www.partow.net/programming/hashfunctions/index.html

Well, at least they were able to test that they messed something up.

There is little reason to use non-crypto hash these days, given fast crypto hash like SipHash.