It makes sense if you think about it. A small startup with 5 Python engineers can’t afford to hire one with heavy experience in Ruby and who knows a thing or two about Python.
First, it may be the case that this person can actually transfer their knowledge and do a good job with Python, but there’s also a chance they can’t and that will disturb the entire operation for a team that small, that’s a big risk.
Also, this person is not going to be paid as a Jr. Python because they have the options to be paid as a Sr. Ruby somewhere else, so the startup need to step up their pay game, now that’s a big and expensive risk.
A big company like Google, on the other hand, can afford to have that risk. If everything goes right, they end up with a great Ruby and Python engineer, if things goes wrong, Google is so big they probably have some other projects somewhere that can use a good Rubyst, and in the worst case scenario, if the person really can’t fit anywhere, a company like Google can still afford to let them go and eat the costs of a wrong hire.
edit: mav3rick has clarified below, it does indeed seem code agnostic.
Provides list of the three languages you specifically need to code in.
Perhaps agnostic means something different in my language :p