[AlbertCory]

Since this is about Google, where I worked:

> The Google spokesperson said that caste discrimination has “no place in our workplace and it’s prohibited in our policies.”

I can testify that Ads was very heavily Indian, much more than the general employee percentages would predict. I don't know if it still is. Sridhar Ramaswamy (https://en.wikipedia.org/wiki/Sridhar_Ramaswamy) probably had something to do with that. From informal observation, IIT was a very common alma mater for them.

Yet, within the SmartASS team in Ads, a huge percentage of the engineers were Canadian. I worked in the office right next door to them for a while.

What does this tell you? People tend to refer their friends and give good recommendations to people they know. A Canadian is more likely to know other Canadians. So is that "discrimination?"

[lozenge]

How is your argument relevant when there were 20 complaints from actual Googlers about caste discrimination? Unless you are actually Indian and aware of the castes you wouldn't have any idea whether caste discrimination was happening to your colleagues.

[AlbertCory]

Did I say there was?

[kelnos]

> So is that "discrimination?"

Yes, absolutely. It may be passive and unconscious, but it's still discrimination. I bet a lot of more-than-qualified candidates -- many likely even more qualified than the people who ended up getting hired -- were rejected.

Unless you're going to assert that Indian people are uniquely the most qualified and best at building advertising systems (which I hope we can agree would be an absurd assertion to make), it's a near certainty that better candidates were passed over.

And looking at it from the potential candidate perspective, I personally wouldn't want to work on a team that is heavily over-represented by any one race (even my own), and I wouldn't be surprised if that "scared away" some great candidates (of other races) for positions as well. This is no different than the archetypal example of a woman engineer being uncomfortable joining a team full of men.

> People tend to refer their friends and give good recommendations to people they know.

I've come to believe that referrals are great when you're a tiny startup trying to find people whose work you can trust (since dead weight can be fatal to a new company), but become less and less useful -- and sometimes even counterproductive -- as a company grows in size.

At any rate, at any non-tiny company, any referral should be put through the same interview process as the other candidates, and should be judged based on the interview, not on the referrer's opinion. The referrer should not even be a part of the interview process, anyway. Not just when it comes to their referral, but (if possible), they shouldn't be a part of the other candidates' interview panels either, as they may be (at best) unconsciously biased against the others.

> A Canadian is more likely to know other Canadians.

Canadians aren't a racial group (and a Canadian may be white, black, native, Asian, Indian, whatever), so I don't think this particular example has anything to do with the rest of what you're talking about. If that team mostly or exclusively comprised white Canadians (or Canadians of any other single race), then yeah, maybe there's an issue there. And regardless, a team comprised of the same $X -- where $X is pretty much anything -- should be a red flag. To me, that's a sign that the team may be cliquish and discriminate (even unconsciously) against anyone who might join the team but be an "outsider" from the perspective of $X.

[AlbertCory]

> Canadians aren't a racial group (and a Canadian may be white, black, native, Asian, Indian, whatever), so I don't think this particular example has anything to do with the rest of what you're talking about.

They're not, but it has absolutely everything to do with it. As you agreed (I think): People tend to refer their friends. Canadians often went to the same schools (Waterloo, McGill, UBC) and if you asked one about another one, quite possibly they'd know someone who knew him or her.

[sleepybrett]

If I ended up on a team that was mostly canadian I would assume that the team was originally built in an office in canada.

[AlbertCory]

AFAIK it was not.

[camgunz]

It's at least a lack of diversity processes, metrics, and benchmarks, which leads to discriminatory outcomes. Or in other words, someone should've said, "hey, this team is X% Indian, maybe branch out a little". It's the exact same dynamic that leads to heavily white/male teams, "hey I made all my friends at institutions that didn't at all strive for diversity... and that's where I hire from... hmm."

[za3faran]

I wonder how that sits with the diversity and inclusion directives that are everywhere now. Maybe they turn a blind eye when it suits them, or use other teams to offset the difference.

[jimmydddd]

From what I've seen from the outside, it looks like US companies focus on what they view as "non-core" departments to reach their overall diversity goals. So, for example, in a software company, the coders might largely be of one or two ethnicities, and then the company's diversity goals are met by hiring diverse candidates in, say, HR, legal, payroll, etc.

[AlbertCory]

idk, but there's always a question of how big a unit you look at when computing ethnic percentages. Does a group of 10 have to have 5 women, 3 POC's, and one LGBT? Or can you measure DEI only for very large groups, like the whole company?

Probably for the government, it's the entire company, or maybe each large strata of it.

[kelnos]

For a team of 10 I absolutely agree that you're probably not going to get that fine-grained, and you'll probably look at larger groups of 100 or more.

But it still can be useful to spot-check smaller teams. If that team of 10 is, say, 100% male, or 100% Chinese, maybe that's something that deserves a second look. Not with guns-blazing, "you all are obviously sexist/racist", but... a second look, nonetheless.

[tekla]

What if it was 100% women, or 100% Black?

[cududa]

What if they were 100% Gelgameks? It's a nonsense question because at any large tech co, there aren't any engineering teams that are 100% women or Black. I don't believe that's a reflection the skills of race or gender lacking, just that there aren't any tech co execs that would feel comfortable with that due to their own bias and overall corporate culture..

[AlbertCory]

> What if they were 100% Gelgameks?

that's the nonsense question. This isn't 100%, but there were Google managers who prided themselves on having very disproportionately large numbers of black or female employees. SRE groups, especially.

In addition, there are plenty of teams that are almost all Chinese. Or Indian. Or, of course, white male.

[LordShredda]

It is discrimination depending on why they were recommended. That's the hard part of stopping discrimination, it's never found until it's obvious

[AlbertCory]

> depending on why they were recommended

so if the "why" is "I worked with this person" that's bad? Or not?