4 ms·
I think it's mostly because the statistics don't tell the story. I've been on the hiring side for 3 companies. We'd take the first non-idiotic person who applie
by muzani 7y ago
I think it's mostly because the statistics don't tell the story. I've been on the hiring side for 3 companies. We'd take the first non-idiotic person who applies.
If you're the kind of person to be a friend, you'd likely be hired. If you're the kind of stranger looking for a job, you'd be unlikely to be hired. If you're the kind of person to hang out on HN or some Facebook/Slack meme groups, we couldn't afford your salary.
Out of that, it would be roughly 90% of applicants who don't qualify, 5% who we can't afford, and somewhere in between which we just settle for. Many new graduates can't do FizzBuzz, many people who apply for a senior position can't reverse a linked list.
- keenmaster 7y ago“We'd take the first non-idiotic person who applies.” I think it would still help for a company like that to display stats. For example, if you’re a qualified candidate who is applying early, you’d see an abnormally high probability of callback. If you’re an unqualified candidate, you’d see a very low probability of callback. I mentioned that ML can help calculate these probabilities. In addition, I think it might help for companies to secretly encode “hard” job requirements. Most requirements are in fact soft, but the applicant never knows which is which. If they don’t meet a hard requirement, they should see a 0% probability of callback. Full stop. ML would do a good job of assigning importance weights to each of the requirements which aren’t encoded as “hard,” but it would start from weights provided by the employer when there isn’t enough data.
- tom_b 7y agoBut stats like that also can signal being over-selective. I would gently suggest that it is a mistake to think in terms of unqualified vs highly-qualified candidates and ML for "callback probability". I don't want a good, but not FANG-level, candidate to NOT apply for our standard software engineer opening just because 74 people already did. Why? We had 60+ candidates for a recent posting for a junior developer role with some experience - think 1-2 years or some good school projects. Somewhere around 45-50 of those applicants were easily HARD-FLAGGED don't interview at all after a quick cover-letter and resume review by a technical person using a rubric. That rubric was scoped only to weed out applicants with absolutely no meaningful experience. While I also hate the laundry-list approach for job listings (Must Have: Expert Java and C#, Spring ORM and ActiveRecord skills with deep understanding of React internals), I think the best way to attack that is with a short, but focused cover letter. That should include a short paragraph or two about a candidate's hands-on experience and speculate how their experience might apply to the position. I guess what I am driving at is that this idea of ML assigning importance weights to "soft" vs "hard" requirements is, at best, just another low-value signal in job postings that are already full of low-value signal.
- keenmaster 7y agoJust like on your LinkedIn profile, there would be way more than a simple list of keywords and skills. The keywords would be part of a full profile which includes universities/degrees, certificates, years of experience, accomplishments, etc...some of which are easier for an ML to train on than others, but my point is those are stronger signals than “Java: Yes.” If there is residual fuzziness in the process, the platform can add skill tests (taken at a test center) that can get listed on your profile. A third party ecosystem of skill tests could emerge, some of which would be more prestigious/respectable than others. Some would have a binary output, and some would give you a score that is benchmarked against the population of test takers. This would only really be necessary for more technical skills. I’m not talking about rinky dinky khaki cubicle test centers. I’m talking about new age corporations with $1B market cap that develop an unparalleled ability to assess skills so that corporate HR departments don’t have to. These testing companies would be part of the ecosystem developed by the theoretical ideal, statistically-oriented, applicant friendly job platform that I’m talking about.
- muzani 7y agoWell, the thing is that interview calls are based almost entirely on resume and cover letter quality, not so much the skills. A hard requirement is that they should understand how case-sensitive and punctuation works, and many applicants can't even get their name in proper Name Case. Feedback is, well, if you school for 11 years and still can't use grammar where it matters, I'm not sure it helps. But too many do this. We did reject one smart guy for an Angular post, because his only technical skill was WordPress and the gap for that is just too big. But we interviewed people with experience with only PHP and no JS. Skills are learnable. There's also weird stuff like rejecting people because their expected salary is too low. It's a pattern that those with low salary expectations performed worse technically than juniors. 0% callback is also rare because of that. I got my first, web dev job, with no experience in anything other than HTML and an EE degree. Rejection rates for a role are 100% until the first hire.
- tom_b 7y agoMatches my experience on the hiring side. In the group between "don't qualify" and "can't afford", we have a hard time with candidates who are unable (or maybe lack the communication skills) to effectively talk about how their experience+potential are a good match for even a generic job opening. We don't do any algorithmic/challenge screening during interviews. A common interview fail for us is to have a candidate that can't generalize from or build on something from his/her resume. This usually comes out during "tell us about project X that you worked on" and comments from our side like "oh, that sounds like project Y here. But Y is different this way - what are your thoughts about how to handle <insert small-scoped, specific difference>?"
- keenmaster 7y agoStats would be displayed prior to getting any interviews. If someone gets an interview and then fails at it, that’s on them. Stats can still help with baseline interview performance though. Machine learning algorithms can find a correlation between experience (or lack thereof) and interview failure. In response, the algos will encourage those candidates to apply somewhere where they have a better chance of getting in. However, those candidates can supplement their skill set. They can take tangible steps to increase their likelihood of getting the job. Let’s say the applicant is a senior engineer but you’re afraid they aren’t keeping up with the latest frameworks. And let’s say that the employer (or the ML also) assigned high importance weights to those frameworks. Here’s what you’d see happen: - The senior engineer with 15 years of experience at several BigCos but doesn’t have experience in that framework will see a low probability of callback - The senior engineer with 10 years of experience and only one BigCo who’s skilled with that framework (the right keywords, and 2 MIT edX certificates in it) will see a high probability of callback. What’s exciting is that this will encourage people to acquire exactly the skills that will help them get the job that they want. The applicants would be able to do scenario analysis on the job platform and see how much better their chances would be if they acquire a certain skill set. This would allow them to tune out the noise and focus on what it takes to get the type of job they want. Of course, the more senior the applicant is, the more accomplishments at past jobs matter. Even then though, senior applicants usually only display a fraction of their past experience. The platform can show their callback probability change in real time if they add a meaningful past experience or accomplishment. The possibilities are endless. Are you considering doing a XY boot camp because you think it might get you into Z Co? Think again, with your experience + that boot camp, you will have a 2% chance of getting into Z Co and 5% chance of getting a similar role with similar pay. That alone would improve the educational and vocational pipeline.
- Jemaclus 7y agoSort of off-topic, but I'm interested in how often you reverse a linked list at your job. Is that something that you expect people to be able to do? If so, what does it tell you about them? If they don't know what a linked list is or how to reverse it, what does that tell you about their real-world coding knowledge or lack thereof?
- Jemaclus 7y agoFor anyone downvoting, it's a serious question. We're debating these kinds of algorithms as interview questions at my job, and I was curious what this brings to the table versus, say, building a CRUD API or constructing SQL queries or anything else. For example, I've been in the industry for 15 years and I've never had to construct a linked list (much less reverse one) in my career. I have had to build APIs, construct DB queries, create distributed systems, etc, etc. So I naturally wonder why constructing a linked list might result in rejection for a job, unless it's something you're actually expected to do on the job for that role. Am I missing something?
- muzani 7y agoWe don't use linked lists, but it's a test to see if they understand pointers. Pointers are used quite often for a number of things. A binary search might be a better test, but it's even harder and not as guided as a linked list modification test. Oh, to note, it's a code modification test, not a whiteboard algorithm thing.
- Jemaclus 7y agoAh, gotcha. That makes a lot more sense. If you were asking them to write a linked list implementation from scratch, I'd have a lot more questions. Having an already-implemented LL and modifying that is a lot more reasonable. Thanks!