by pixelcort on 11/28/18, 4:51 AM with 245 comments
by carlmr on 11/28/18, 2:19 PM
They wanted a 100% candidate. Looking through the list of the things they listed as must haves and that they wanted an internal candidate I showed them it is statistically very unlikely they would even find a better candidate for this position.
The position had been open for almost a year. I think there's a reason. These are probably the same people that go to the press and whine about a shortage of engineers.
EDIT: Made a small mistake there. Anyway I only mentioned it after the interview was over. Not during.
by paultopia on 11/28/18, 7:02 PM
First, we randomly sampled 6,348 applications for 668 different users from TalentWorks. Then we extracted the qualifications from the original job postings and the users’ submitted resumes using proprietary algorithms. Finally, we grouped the results based on qualification match and regressed the interview rate using a Bagging ensemble of Random Forest regressors.
This is... not plausible. Effectively, they're trying to infer causality here, not merely do prediction. That has to be the case, because this is presented as useful advice---to go ahead and apply even if you don't meet all the listed qualifications. But when you're trying to infer causality you're doing social science, not data science, and that means you need to worry about omitted variables.
Here's an example: what if less qualified people who nonetheless apply are more confident. And what if that confidence is associated with other good things that show up on resumes, like attending prestigious schools, having had prestigious prior jobs, having a record of success in some other fashion, or even just doing things like paying careful attention to formatting?
This is why social scientists use tried and true techniques like old fashioned OLS regression with control variables rather than throw everything into a random forest and see if the hypothesized association standing on its own predicts things.
(Insert remark about how companies should be hiring data science people with science backgrounds rather than just pure cs backgrounds here)
by rossdavidh on 11/28/18, 2:47 PM
However, typically, I see 20-30 requirements, many of which (e.g. "passionate about software development", "good communicator") aren't specific enough to tell the candidate whether they are a good match or not (do poor communicators know they are poor communicators?). Of the rest, really only 2 or 3 are actual requirements, and the candidate has to guess which those are.
by mszcz on 11/28/18, 1:22 PM
I'm not a fan of HR in general and it's frustrating that they are often the clueless gatekeepers.
by scarejunba on 11/28/18, 12:41 PM
And let’s be honest having spent “5 years of Python” at a Google is a different story from having spent “5 years of Python” at an Infosys.
To be honest, I’ve never seen a job description that exactly lays out what you’re going to be doing. And the reason for that is that hiring software engineers, you don’t actually know. You want them to also come up with what the future looks like.
by dexen on 11/28/18, 10:36 AM
>I'm not sure why companies over-list [job requirements].
It's not companies, it's the individuals.
Consider this: in a corporate environment, a person that is responsible for hiring but that is not a stakeholder in the success of any particular project, is incentivized to prove that:
- she or he made an effort ("I've posted N ads on top ten websites")
- she or he didn't cause any particularly bad hires
The first incentive favors cookie-cutter hiring requirement lists and ads, in the "nobody ever gets fired for buying IBM" sense. Copy-paste an ad from a different project, adjust a few minor points, file it away.The second incentive favors over-specifying requirements, in the hopes that no particularly bad hire will be made and then blamed on the requirements / ad author.
Suppose for a second a hiring manager or HR specialist were told by project stakeholders "certification X and skill Y are requirements", but figured out they aren't actually key to success - perhaps learning on the job would work out just fine in this case. So our brave hiring manager or HR specialist puts the certification and the skill in the "nice to have" section instead. Now suppose a candidate hired without the certification or skill ended up disappointing and underachieving. The manager or HR specialist would shoulder the blame for not filtering the hires well enough. Thus they play it safe and over-specify.
It doesn't help that there's a persistent, lingering narrative[1] in the press that pretty much all the skilled specialists are in high demand and in very short supply on the job market. This provides a cover for anybody who failed to attract candidates due to over-specified requirements - "the specialists are in short supply anyway".
Source: having been doing guerilla-style hiring with carefully redacted ads for a long while, with repeatably good results.
--
[1] the narrative seems mostly created by the prospective employers in hopes of driving the worker supply up, and thus prevailing wages down
by tombert on 11/28/18, 3:17 PM
Sadly, it does seem like the research-oriented jobs aren't BSing about their requirements; when I've tried to apply to MS Research and the like, they've always declined me due to lack of credentials, since the postings usually require at least a masters, preferably a PhD.
by GhostVII on 11/28/18, 2:03 PM
by bradleyjg on 11/28/18, 2:37 PM
However the essence of the posting was clearly for a programmer. Significantly more than half the people that applied were not programmers. We saw some resumes from seemingly quite accomplished statisticians but we were not looking for a statistician (or “data scientist”).
So I think which 50% matters. If someone is a self starter and has great communication skills but can’t program she isn’t going to be hired as a programmer.
by GuB-42 on 11/28/18, 1:36 PM
by scaleout1 on 11/28/18, 5:28 PM
by zaidf on 11/28/18, 9:41 PM
- super passionate (paying customer for while) about the product
- great designer with a few years of experience
- degree in CS from a reputed school
- picked up react recently
- few years of work experience doing design and coding
He got turned down by the startup with the excuse that they would like someone more senior who has worked on a product like Facebook. I could just smh because from everything I know, if you are senior and worked at fb, you are unlikely to be (or want to be!) an AMAZING designer AND front-end dev.
by hocuspocus on 11/28/18, 1:31 PM
That's good but please, if you aren't Google or Facebook, don't take it to the other extreme. If I have no idea about your tech stack I'm not going to apply, and the same probably holds true for a lot of experienced developers.
by sxp62000 on 11/28/18, 11:46 PM
If you're looking for a job, just keep applying and don't stop till you actually have an offer letter in hand. Don't wait for the employer or recruiter to send you feedback after a promising interview. There are SO MANY reasons a company might decide to go for someone else. For example, culture fit, which sometimes means "will this person stop everything and play foosball with us in the middle of the day?"
by gefh on 11/28/18, 7:03 PM
Doesn't seem to have stopped a lot of people confirming their anecdotes though.
by BenjaminBlair on 11/28/18, 2:53 PM
by codingdave on 11/28/18, 2:53 PM
But I'm not sure that the sales pitch this article is trying to make is valid. They claim at the end that they'll get you 5.8x more interviews. But is that good? In my 3 interview processes, they each brought to light reasons why it was not a match in one or both directions. The article even touched on that, as candidates will self-screen out of jobs that aren't quite right. It would have been a huge waste of time to expand that to 10-12 interview processes that were not matches. Especially when I didn't end up taking a job even from the ones that went well, and re-joined an old team instead.
The screening that happens in the hiring process can feel frustrating when you really need a job, but I appreciate being screened out when it would not have worked anyway. It saves everyone time, and hopefully puts me in a healthy long-term role that will last for years.
by proxygeek on 11/28/18, 6:59 PM
It might be interesting to do a basic pre-processing of the JD to remove all such stop-phrases before evaluating the role being offered.
by bitwize on 11/28/18, 1:56 PM
Not that I apply for jobs to get jobs, not to get interviews. Getting called in to interview means next to nothing.
by jkingsbery on 11/28/18, 2:03 PM
by DEADBEEFC0FFEE on 11/29/18, 7:17 PM
If I can find a 5, 4, 4, 3, 2 of combination of that, I'm happy. That's about 70%. That sort of profile should also provide learning opportunities for the candidate and typical that is intrinsically motivating for technical types.
by TheAdamist on 11/28/18, 12:32 PM
by rb808 on 11/28/18, 3:14 PM
by create_novelty on 11/28/18, 5:18 PM
by mathattack on 11/28/18, 2:10 PM
by pacuna on 11/28/18, 1:31 PM
by MattHeard on 11/28/18, 1:20 PM
by austincheney on 11/28/18, 12:41 PM