4 ms·
I think that an important factor is the extent to which interviewers' assumptions and criteria are aligned. At one of my previous jobs, I noticed the following
by MathMonkeyMan 5y ago
I think that an important factor is the extent to which interviewers' assumptions and criteria are aligned.
At one of my previous jobs, I noticed the following in hiring:
- The senior engineering recruiter, in HR, thought that there was one standard for whether to make an offer, and that team selection happened afterward. The engineering managers who actually had the power to say yes or no instead thought that there was a hierarchy of quality among the positions offered, e.g. "we won't hire him, but send him over to those morons maybe they can use him."
- Engineering managers did not value open source contributions or non-work projects. I pointed out a particularly impressive piece of C++ on a candidate's github, and was told "how do you know that he didn't just get someone else to do that for him?" The manager cared more about the minutiae of undefined behavior in the iostreams library.
- Interviewers are not willing to take the time and energy it would take to grok someone's code and assess what it says about their value as an employee. The result would likely be subjective anyway. It's not unreasonable to conclude that fair assessment is too expensive. It also makes me wonder whether organizations with excellent code review culture are also excellent at interviewing.
Let's keep giving candidates programming-centered IQ tests, preferring people like ourselves, and occasionally when we wish to signal an interest in balanced racial or gender demographics, a committee will quietly adjust offers to compensate.
- theamk 5y agoI am confused about last sentence -- are you saying that requiring high-quality open source code will somehow help with "balanced racial or gender demographics"? Because I am pretty sure this is false. Anecdotally, the women in my group do not have any public code. And web searches found confirmations for this as well [0] [1] So if you want diversity, stay away from methods like the one described in this post. [0] https://www.toptal.com/open-source/is-open-source-open-to-women https://www.toptal.com/open-source/is-open-source-open-to-wo... [1] https://www.ashedryden.com/blog/the-ethics-of-unpaid-labor-and-the-oss-community https://www.ashedryden.com/blog/the-ethics-of-unpaid-labor-a...
- nivenkos 5y agoWhy not? They have the freedom and opportunity to create FOSS code just the same as anyone else.
- AnimalMuppet 5y ago1. Just generally, it's rather exploitive to require something unpaid that people do in their spare time in order for them to get a job. 2. Specifically with respect to race and gender, don't use something to guide hiring that you know skews one way. And don't excuse it by saying "but those other people had the freedom and opportunity". They did, but it still skews, and you still know it does. Don't use signals that are known to be biased.
- nivenkos 5y agoUnfortunately most take-home tests hit those two points too though.
- AnimalMuppet 5y agoThey at least hit the "unpaid work in order to get a job" exploitive part. I'm not sure about the "latent discrimination part". But yes, I regard unpaid take-home tests as filtering for workers who are willing to be exploited...
- MathMonkeyMan 5y agoMy last sentence was not relevant to the topic of the article. It was my jaded summary of how much tech hiring I've seen works. Maybe an edgy non sequitur that didn't belong.