Google is a discrimination factory, but in this case, there's a deeper problem, and its, what I'll call, the "MIT culture".
You have a bunch of people who busted their ass off to go through MIT/CMU/CalTech/Whatever, to learn all those algorithms, the computer science core, etc, and are thrown in the real world where, while VERY useful, are only a small subsets of things that matter.
Then you ask these people, who spent 4 (or 6, or more) years being drilled that the only shit that matters was what they learnt in school, and worked REALLY hard to absorb that, to interview.
What do you think will happen?
You end up with an interview process that, regardless of the actual work, the further away from school you are (ie: the older you are), the less likely you are to pass the interview, give or take people who worked as data or algorithm scientists in the recent past.
Net result: you have a very high percentage of college hire, and your lateral hires will always lean toward the younger side. Any skill that come with experience is almost never tested in interviews to counterbalance it.