"In my wife's case, the problem was that the committee deadlocked on how to level her. She had 5+ years of academic experience, and 5+ years of research experience (think national lab type stuff; research w/o teaching). However, she'd only briefly worked in industry many years ago between her MS and PhD. So rather than hire her, they told her to "go work someplace else for a year, then apply again". (direct quote, as relayed by her recruiter to me, as I referred her)"
(note, the below are all very very much personal opinions :P
Most of what i'm going to talk about is all covered by a large number of books by various people, etc at this point.)
First, sorry to hear about your experience. Hope you don't take the below personally.
Having sat on google hiring committees (2-3 a week sometimes) for the better part of 11 years now, i'm going to say this is probably not accurate. Note that i only do SWE committees, though.
I'm also not sure how you derived this view of what occurred (if you are going by what the recruiter said, know that most recruiters have a jaded view of committees who reject their candidates. Otherwise, i'd love to know what facts you have) . You seem to be assuming a lot of the reasons the committee said what it did.
The answer you received is, in my view, usually the answer given when "we don't think they did well enough to hire them, but think that they may get there with practice". For example, in SWE world, this is usually what happens when committees feel like these people are on the wrong side of being able to design and implement stuff, but saw things in the interviews that made folks believe they may be able to get above that bar.
(or put bluntly, not obviously worth the risk right now, may be worth the risk in the future). That isn't to say Google doesn't hire people who need mentoring in coding/etc, it's more "committee didn't feel it was a sure enough thing to take the risk".
Note that there are second chance, etc committees if the recruiter felt the committee was wrong. Committees aren't always right, and they definitely aren't meant to be always right.
The leveling decision is also completely independent of the hiring decision in just about all cases. There's some interesting history here, but the statement "both decides if you passed your interview, and then decides on your job level", is wrong now, and was wrong for pretty much everywhere but MTV for a long time.
In any case, if what you say occurred, generally in SWE world the only reason they wouldn't make an offer at the lower level is if "they've not grown in 10000 years to a sane level of engineer, so we don't ever expect them to", or the person has said they wouldn't take it and so it would be pointless. In fact, most of the discussion that used to happen was "we think they are qualified for this lower level, is it worth discussion or have they said they wouldn't accept it" (because they have competing offers.
The first seems ... not the case here, unless your wife did very badly (IE would have had to be two levels down). Sorry. So i find your experience very very strange, though i don't do data scientist committees.
You started with:
"Google often does not know how to hire senior people who are not luminaries, this screws people in their 40s."
This is not right.
IMHO, Google often tries not to hire people who have failed to grow to the level and skill set they want out of people with a large amount of experience or who want a senior position.
But that's literally not about age (despite people wanting to claim it is), it's about growth vs experience. They do even try to account for what the experience was, etc. But usually it's the other way around - person wants senior position at Google, isn't qualified for it, complains google hates people who are older. Google would, in most cases, happily offer them a position they are qualified for, they just don't want that one (or google doesn't have such a need ATM). Personally, I wish that being XXX level engineer at some company mattered. But i've also met plenty of people who got promoted a lot just for surviving long enough, and couldn't code/design their way out of a paper bag despite being "senior principal engineer" or whatever. So i understand why they use interviews to try to gauge level instead.
Past that, if you've spent 35 years in industry, and can't code/do algorithms/design/or any measurable metric better than a person fresh out of college, and aren't more self directed, etc, than the fresh out, it may not make sense to hire you.
But again, that's still not about age.
Because the person fresh out of college could be 70. I'd still say "hire that person", because they still have potential. The person who has 35 years of experience and still not grown is a known quantity, and that known quantity is not above the bar.