Programming thread

  • 🔎 If you see an alert for analytics.usips.org, this is a self-hosted system I am running as a test. It is GDPR compliant and collects no invasive information.
  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Touring machines
26 Jaguar E Type RM Sothebys.jpg
Brother's idea of a computer
 
I've been on the company side of finding dev candidates the last few weeks and our HR is swamped with application submissions from bots and generally unsuitable candidates.

Obviously they didn't go through each applicant in the finest detail before sending them to us(a few devs who stepped up to handle the technical interview), but only a handful of those sent through the filters could even hold a technical conversation about stuff they specifically put on their resume.

It was really disappointing after how shit the market has been for devs to see how these people shit up the pipelines. I definitely think companies are doing their part to shit up the market but these fake it till you make it devs aren't helping and have been annoying to me personally.
 
I've been on the company side of finding dev candidates the last few weeks and our HR is swamped with application submissions from bots and generally unsuitable candidates.

Obviously they didn't go through each applicant in the finest detail before sending them to us(a few devs who stepped up to handle the technical interview), but only a handful of those sent through the filters could even hold a technical conversation about stuff they specifically put on their resume.

It was really disappointing after how shit the market has been for devs to see how these people shit up the pipelines. I definitely think companies are doing their part to shit up the market but these fake it till you make it devs aren't helping and have been annoying to me personally.
I'm curious how you are assessing them from a technical perspective. What's your approach? I've been in a bunch, and straight up froze up when describing some aspects of my work.
 
I'm curious how you are assessing them from a technical perspective. What's your approach? I've been in a bunch, and straight up froze up when describing some aspects of my work.
I've done a ton of interviews. Given and received. I WILL ask you questions you can't answer. But you should be able to form a proper hypothesis, ask questions and at least tell me how you'd solve it. Things ranging from "tell me what a linked list" is, well, let's see, it's a list, and it's linked... etc. I do more sysadmin, so a standard one was "Tell me how a computer boots" which instantly told me the level of depth I was going to be working with.
If I ask about former projects I want to know what the problem or task was, how did you evaluate them, how did you solve them, what tools you used, was the customer or boss happy, etc.

On the other hand I say I will ask questions you don't know the answer to, I HATE stump the chump style stuff. "Tell me how to implement a red-black tree in native 6502 assembler." Yea, you not knowing that tells me nothing about what you do know. As I said, my "you don't know" questions are to probe your knowledge AND problem solving skills not just showing off that I know how to spell C.
 
I'm curious how you are assessing them from a technical perspective. What's your approach? I've been in a bunch, and straight up froze up when describing some aspects of my work.
I find that I can sus out maybe 80-90% of what I need to understand a candidates ability by simply asking them pointed questions about the items they put on their resume. It sounds simple but it has been that simple in my experience. Start general and drill down for more detail.

"I secured my company's APIs preventing unauthorized access to internal APIs"
What did you use for authentication? So you received JWTs? How did you get the permissions from it? How did you separate out the various steps of handling a request from the HTTP layer to the data layer? Did you run into any issues with the way you handled it? For higher level positions you can get into bigger questions that show they have expertise in the area. Did you own the identity platform or did you use a service/have a team for that and why? How do you establish trust with the IDP and between API servers? Did you collect metrics, what did you find important to track? Was there any auditing, how did you make it accessible and actionable? What kind of monitoring did you have in place and what would warant a key request? Did you have any long running actions based on requests? Maybe they mention an event system and I ask how they model events, when they extend an event with a new field vs making a new event, if it was idempotent and how.

I'm pretty bad at interviews myself due to anxiety, but I could answer these types of questions. It ends up being conversational which tends to put the candidate at ease and the details flow a little better being able to have some back and forth.
 
Last edited:
I've been on the company side of finding dev candidates the last few weeks and our HR is swamped with application submissions from bots and generally unsuitable candidates.

Obviously they didn't go through each applicant in the finest detail before sending them to us(a few devs who stepped up to handle the technical interview), but only a handful of those sent through the filters could even hold a technical conversation about stuff they specifically put on their resume.

It was really disappointing after how shit the market has been for devs to see how these people shit up the pipelines. I definitely think companies are doing their part to shit up the market but these fake it till you make it devs aren't helping and have been annoying to me personally.
Is there any approach in the industry to address this? My naive idea to address this would be to have the candidates be pre-interviewed by AI. Would still be a cat and mouse game against people trying to apply with AI in turn of cause, but still prob. less duds then when just filtering by whatever they put in their resume, right?
 
Is there any approach in the industry to address this? My naive idea to address this would be to have the candidates be pre-interviewed by AI. Would still be a cat and mouse game against people trying to apply with AI in turn of cause, but still prob. less duds then when just filtering by whatever they put in their resume, right?
I think that public positions are going the way of the Dino. I have never gotten a decent paying job from a job posting. I think firms just for finding qualified candidates and matching them to hiring managers are going to do very well in the future. I see good recruiters being in higher demand.
 
Is there any approach in the industry to address this? My naive idea to address this would be to have the candidates be pre-interviewed by AI. Would still be a cat and mouse game against people trying to apply with AI in turn of cause, but still prob. less duds then when just filtering by whatever they put in their resume, right?
I stumbled upon a few AI interview services when my engineering manager tasked me with looking for some new coding challenge options. I have no experience with them on either side, so I'd need to build trust in it being able to properly filter bad candidates AND not turn off good candidates. At a glance I was thinking if it was automated as an early step, short in duration and a human could easily go through the transcripts and spot check the videos it could be nice to have. I think this still requires someone in engineering to give final say on the AI's assessment, but it would filter the pool some more and is much less time consuming than sitting down a couple of devs for a full tech interview.

I think that public positions are going the way of the Dino. I have never gotten a decent paying job from a job posting. I think firms just for finding qualified candidates and matching them to hiring managers are going to do very well in the future. I see good recruiters being in higher demand.
I always fall back on networking being your best tool to land a job. Do you find recruiters to be valuable? I've used them before but always ended up in a direct job during my job searches.
 
I stumbled upon a few AI interview services when my engineering manager tasked me with looking for some new coding challenge options. I have no experience with them on either side, so I'd need to build trust in it being able to properly filter bad candidates AND not turn off good candidates. At a glance I was thinking if it was automated as an early step, short in duration and a human could easily go through the transcripts and spot check the videos it could be nice to have. I think this still requires someone in engineering to give final say on the AI's assessment, but it would filter the pool some more and is much less time consuming than sitting down a couple of devs for a full tech interview.

Literally on the last page I have a post moaning about how I gave up pursuing a gig after they wanted me to have a half-hour video call with a clanker. I definitely think that as the length of the screening call increases, there's going to be an exponential increase of sane devs dropping off and only the truly desperate staying. Hell, there's probably going to be a decent number who opt out at the point there's such a call regardless of the length.

Please use these things judiciously. I get that you need to filter out the mouthbreathers, but surely that determination can be made in just a few minutes of talk with a bot.
 
I always fall back on networking being your best tool to land a job. Do you find recruiters to be valuable? I've used them before but always ended up in a direct job during my job searches
Not all recruiters are created equal. A few of them want to actually directly place you with a company. They may actually work for the company in the HR department.

The vast majority of recruiters are trying to get you to join their bodyshop where they use their network to get you a job, and bill the client more in exchange for not having to pay for health insurance or W2 matching. The body shop pays your W2 and you get no PTO or benefits. There's some Capex tax loophole they haven't closed yet, but that doesn't make sense in the long term.

The issue is that most recruiters are divorced from the technical side, and spam LinkedIn in mail or email.
 
Literally on the last page I have a post moaning about how I gave up pursuing a gig after they wanted me to have a half-hour video call with a clanker. I definitely think that as the length of the screening call increases, there's going to be an exponential increase of sane devs dropping off and only the truly desperate staying. Hell, there's probably going to be a decent number who opt out at the point there's such a call regardless of the length.

Please use these things judiciously. I get that you need to filter out the mouthbreathers, but surely that determination can be made in just a few minutes of talk with a bot.

I read your post and I thought the same thing about how it can even be a net benefit, it needs to be quick and easy and even then it will still turn off some number of candidates. Overall it's a balancing act and adapting to the times the best we can. Think about it this way, if we don't utilize AI at all and try to be as human as possible, you'd still have a difficult time getting to the interview stage with a fitting resume, so this comes from a place of getting the reasonable candidates into view in the first place.

Not all recruiters are created equal. A few of them want to actually directly place you with a company. They may actually work for the company in the HR department.

The vast majority of recruiters are trying to get you to join their bodyshop where they use their network to get you a job, and bill the client more in exchange for not having to pay for health insurance or W2 matching. The body shop pays your W2 and you get no PTO or benefits. There's some Capex tax loophole they haven't closed yet, but that doesn't make sense in the long term.

The issue is that most recruiters are divorced from the technical side, and spam LinkedIn in mail or email.

I feel that. I've only ever used the kind that link you up with a company for payment/percentage. I ignore all linked in recruiter messages but I did have a coworker switch jobs because a recruiter reached out to him on linked in and got him into a better job. Might've been an internal recruiter instead of a recruiting company though.
 
I stumbled upon a few AI interview services when my engineering manager tasked me with looking for some new coding challenge options. I have no experience with them on either side, so I'd need to build trust in it being able to properly filter bad candidates AND not turn off good candidates. At a glance I was thinking if it was automated as an early step, short in duration and a human could easily go through the transcripts and spot check the videos it could be nice to have. I think this still requires someone in engineering to give final say on the AI's assessment, but it would filter the pool some more and is much less time consuming than sitting down a couple of devs for a full tech interview.
An AI interview service tells me that you either don't care or are ineffective at hiring the people who will be my coworkers on your own.


It's like an artist who sends you a portfolio of AI generated art when you ask for examples of their work. Can they draw something original? Maybe, but all signs point to someone who can't.
 
An AI interview service tells me that you either don't care or are ineffective at hiring the people who will be my coworkers on your own.


It's like an artist who sends you a portfolio of AI generated art when you ask for examples of their work. Can they draw something original? Maybe, but all signs point to someone who can't.
Of course we're ineffective at finding a suitable applicant in over a thousand applicants for a single role without wasting time on unsuitable candidates and bots, we're human. That is precisely the problem we are trying to solve because we care about who we are hiring. Every person that could've been screened out is one less person in front of you in the queue, so we need to explore our options. No one is saying AI should pick the new hire.
 
Of course we're ineffective at finding a suitable applicant in over a thousand applicants for a single role without wasting time on unsuitable candidates and bots, we're human. That is precisely the problem we are trying to solve because we care about who we are hiring. Every person that could've been screened out is one less person in front of you in the queue, so we need to explore our options. No one is saying AI should pick the new hire.
are false negatives or false positives for choosing who to interview worse?
 
Back
Top Bottom