OP here. I posted this as I'm continually troubled by tech recruitment - I think it's fundamentally broken and has been for years now.
I'm verging on writing a long post on it, but came across this article as it included some simplified data/research on the topic (there isn't much around!).
A lot of (or the vast majority) of recruiters have little concept of the true value that sits behind a resume/CV - they don't even have the capacity to understand/interpret the content at times, but there are a few that aren't as bad as the rest. A hat-tip to them!
I've had to work with a lot of recruiters at times in tech leadership roles and not by choice. I dislike working with them and generally because most of time, they don't really care about fit - they're simply looking to hit their monthly, quarterly or annual target numbers. I've also seen actions by recruiters that verge on slander/defamation and horrors I don't want to see ever again.
Where CV's are concerned, it's a crap shoot and I'll happily admit that I hate writing my own...
In the comments, someone mentioned "hiring direct". Now there's a method that truly does work. I've hired more than 150 devs personally and the vast majority have been through direct hires via network and events.
Whilst it's super important to hire for skills, hire for attitude first. Skills can be improved (and taught if needed). Attitude is everything!
> We strongly encourage anyone looking for work in this market, especially if you come from a non-traditional background, to stop spending energy on applying online, full stop. Instead, reach out to hiring managers
Absolutely true, at least for now. The techniques presented afterwards are straight-up inside sales prospecting 101 I predict that HMs will build up thick firewalls as desperation increases and these techniques become more widely known.
In my experience (and Lazlo Bock's in "Work Rules!") most recruiters are bad, and the good ones are good enough that you should stick with them across their career. Thus in aggregate them performing poorly doesn't surprise me; I'd be curious to see a study on whether the good ones are consistently good or if it varies from test to test. That said I don't think ANYONE is going to be amazing at parsing only resumes - part of the point of having recruiters whether internal or external is to be able to have a screening call with a much larger number of candidates.
The software engineering world is fortunate to have the concept of open source, which allows filtering of candidates based on previous published work. Since today we have many open source active candidates, it makes selection much easier compared to unknown candidates. My advice to job seekers - you’re reducing your employment chances if you dont have example projects published.
A good article. And then below it? Links to the usual leetcode-style questions that have little to do with job performance: invert a left-handed negative tree, etc. How about: figure out why that critical cloud function silently fails 0.1% of the time?
I got the impression that recruiters miss the ability to read. But he says that recruiters actually do try to parse the CV for 31s. There's probably the explanation: parsing VS reading.
Management people do not want to deal with awkward, weird, introvert developers/academics. Instead, they get lower-level people as recruiters and administrators that are like themselves, but willing to deal with developers and such.
Meaning, we did it to ourselves by not knowing how to deal with people.
Developers like to treat all problems as technical problems. And human problems with incompatible incentives typically don't have a technical solution.
Failure to understand this and to develop relevant soft skills is what's causing frustration for a lot of (very talented otherwise) devs.
As a marketer this hits home .. we’re responsible for making sure the devs, who get paid double what we do, can pay their astronomical rents and Uber eats habits
And the level of disrespect we receive in order to do so..
Edit -
I’m no longer in the tech industry and this is a big reason why.
I was never treated poorly per se, but I was always treated as a second class citizen. Salaries, responsibility, seniority .. you begin to realise that there’s a glass ceiling which you’ll struggle to break through.
The breaking point for me was when after conducting a detailed analysis of some data using Looker, an entry level engineer was tasked with making sure my analysis was correct. (I was the most senior marketing employee at a company with a $200m Val and reported to the CRO.)
And then his correction of my analysis - which was accepted unquestioningly - was littered with mistakes.
Once we had gone through it together; he had corrected his report and apologised to me.
Are there any other business constructs that are similarly as universally bad as recruitment? Like which stakeholder actually thinks they're any good? Maybe certain kinds of outsourcing would be comparable, where someone sees a cost savings but otherwise doesn't care at all about the outcome?
They tend to be non-software-engineering companies that decided they want to do a bit of software dev. These jobs can be good in some ways (varied job tasks, exposure to different fields of expertise, good work/life balance) and bad in others (no real software engineering culture, salaries typically lower, limited career growth prospects).
Pretty interesting read as someone looking for an internship and getting rejected by just about everyone. My experience line up pretty well with their observations, though a bit more explicit. They state the fact that they’re competitive and very specifically that they want “Cambridge-level” students. I doubt they even looked at the rest of my portfolio.
> would AI do better
Oh please don’t. I would rather be rejected by a fallible human than some paper clip maximizing algorithm that holds no responsibility
I’ve had gitroll.io used on me before and it failed to detect any valid repository on my GitHub profile, had to manually show them my profile but at that point, they were already suspicious/doubtful. There is no way to tell what the AI hallucinated unless it’s outrageous and you get told
Well, that’s disturbing.