Interviewing a software engineer who prepared with AI

Too real. I once got turned down by the Apple Store for a retail position because I wore a collared shirt to the interview (after being told in advance not to wear anything formal). Interviewer let me know I came off as too formally dressed to get their vibe. The discrimination/bias was real.

I didn't wear a suit, but in 2012 I wore slacks, tucked in collared shirt, and a tie, and got the same response from Microsoft. It was for an internship which is hilarious.

I interviewed elsewhere and one other time I wore an Oxford. I passed the university interview but the hiring manager told me for the on campus interview to not wear that again, or I'll stick out too much. I wore a plain T-shirt and have been happily employed for 10 years here :)

[in NYC accent] What, you think you're better than us

I was hired by MS wearing a suit in the 2000s

So tech got rid of the suits, but kept the desire to judge people at interviews based on their clothes?

Great. Fantastic job everyone /S

Very close story here as well. lol. “You can always underdress but never overdress!”

Thanks mom!

To be fair, "suit up" usually means to put on a uniform rather than to wear a suit. The phrase seems to have originated in sports. T-shirts and hoodies are the uniform of tech.

"Be yourself"? What does that mean?

What if wearing a suit is "being myself"? You'll be penalized in tech for that.

Not everyone views the wearing of suits as some kind of punishment.

As I learned, you can also be yourself, never wear a suit on the job and still wear one for the interview. First impressions count. Once people know I can wear a suit they just don't seem to mind me in shorts anymore.. So I might have a social skill after all :D

It has spread to other industries and circumstances other than suits.

I now have no idea how I am supposed to dress for most things other than formal occasions like wedding, funerals, or formal dinners.

Seeing someone wearing a suit for a dev interview would make me think one of the following:

(1) This person really really needs the job. Probably is in a bad negotiation position, due to this urgent need.

(2) Are you here to impress people with looks, or with your skills?

(3) They take looks way more serious than they should, maybe not focussing enough on the technical side of things.

(4) Hopefully this is not an "EnTeRpRiSe software" developer, and if they are, hopefully they don't work on my team and if they are, hopefully my next up manager does not get blinded by fancy clothes, instead of technical reasoning.

That said, I would try to keep an open mind about the person, but they would be initially sorted into the category of managerial or close to management, rather than close to the other engineers, which is not a positive signal to send.

It's not unusual in Europe - but then Europeans tend to dress smarter than Americans in general too.

The general rule is to dress one step up from those in the role. Everyone in hoodie and shorts? Wear pants. Everyone in collared polos? Go business casual with maybe a blazer. Showing up a level lower makes you look unprepared. Showing up some levels higher, like in a suit to a hoodie shop, shows lack of research and reading a room.

In start ups, I have seen candidates nearly rejected just on a suit alone. Def started them on the wrong foot impression wise.

is it bad form to just like, ask your HR screener what the general dress code/vibe is like?

I'm curious about this. When I've hired, I've always wondered how I can actually tell (a) what soft skills are required for the role and (b) whether a candidate has them.

People sometimes think that's a silly thing to ponder: it's obviously obvious! But at most places I've worked, we spend lots of time defining the technical skills required for a job and handwave the rest.

I guess people assume "they'll know it when they see it". But there's a lot of ambiguity. Parent comment suggests that being comfortable sitting in a conference room for an hour is an important part of their job. In some workplaces that would be an odd requirement. I've worked at places where the important thing was being able to go away and make progress on something for a few weeks.

I suspect there are people with autism reading these threads and feeling disheartened. It would be easy to leave with the impression that neurotypical people expect you to make all the effort and they won't try to meet you half way. Some workplaces are like that. But in all the talk about neurotypical vs neurodivergent, it's easy to forget that neurotypical people are a varied lot, just like neurodivergent people. Workplaces are a varied lot too.

> Hiring is all about finding the best candidate. If you find you cannot function sitting in conference room with three other people for an hour, there is a 100% chance there is a better candidate suited for the role, even if his/her technical skills are less than yours.

> Jobs have soft skill requirements, and there is nothing bigoted about that.

Everything you just said also applies to someone who's deaf, blind, or physically impaired.

Apply that same logic to someone with one of those conditions, and enjoy losing the discrimination lawsuit.

There is a big difference between being in a conference room for an interview where you are judged, and on a regular work day. There is for me, and I'm old and have done dozens and dozens of interviews, largely successfully. Don't summarily judge people, especially if they're not neurotypical, as often happens in software.

> If you find you cannot function sitting in conference room with three other people for an hour, there is a 100% chance there is a better candidate suited for the role, even if his/her technical skills are less than yours.

This assumes that was the job? What if the job never talks or sits in a room with anyone?

> Hiring is all about finding the best candidate.

Then what is leet code about?

there is a world of difference between interacting with three people you don't know for an hour for the explicit purpose of stress testing your experience and knowledge and interacting with three people that you talk to every day talking about a project that is well familiar to you.

[flagged]

I'm deaf and rely on real-time captions for calls. In an in-person interview scenario, I'm at a huge disadvantage and not able to perform at my best. In a video call, I'm on equal ground.

It's not as simple as "requiring people to be able to interact with other people."

No, but ignoring their disabilities and saying to just learn to do better is not good enough. It is no different from telling a deaf person to learn to hear with their other sense, it doesn't make sense as their disability is what prevents it. People do need to be able to interact with other people, it just doesn't work like it does with non-neurodivergent people. It takes an effort on both sides. Quit putting it all on the person who cannot do what you want. That is the bigotry.

> they will of course favor those who they are able to effectively interact with. This is not bigotry, is it?

If "those who they are able to effectively interact with" ends up meaning only people who look, act, or believe like them, then yes it absolutely is.

So fly them to multiple destinations? I was hired 1 year ago and interviewed with ~14 people all living in different locations. That could be paired down, but it won't ever reach the single destination that the OP is referring to.

You can always rent a conference room for an hour or so somewhere in between.

You're merely reinforcing their point. Its so out of fashion it would be considered a bold or even edgy choice just as dressing casually once would.

With all the unemployed tech workers, would it just make sense to hire someone who knows their salt to do recruiting and interviews? Recruiters always seem to have a blast moving between random high-level companies and ghosting people over text, socials and the phone. If they lack both the social skills and the technical knowledge, I don't know what their value proposition is, but compared to chronic underemployment after actually learning Java, C, C++, they're clearly winning.

Interviewing has also become harder too. You try to search the net during the interview, because you forgot the name of a thing, and the interviewer will assume you are running with an AI chat, and are cheating the interview.

I have seen wild things in the last year.

People taking minute-long pauses before answering questions. People confidently saying things that are factually incorrect and not being able to explain why they would say that. People submitting code they don't understand & getting mad when asked why they wrote something that way.

I get that candidates are desperate for jobs, because a bunch of tech companies have given up on building useful software and are betting their entire business on these spam bots instead, but these techniques _do not help_. They just make the interview a waste of time for the candidate and the interviewer alike.

I haven't seen the use of AI in interviewing (non-tech) yet, but something has definitely changed: People are now applying in droves.

Because it's a nonsensical reduction and false equivalence.

It's like if you saw a headline that some grocery stores were price fixing, so you decide it's only fair if you steal from your local grocery store. One bad behavior does not justify another in a different context. Both are wrong. It's also nonsensical to try to punish your local grocery store for perceived wrongs of other grocery stores.

That's why it's such a ridiculous claim: Two wrongs don't make a right and you don't even know if the people you're interviewing with are the same as the people doing the thing you don't like.

Using AI to review and improve your CV would make sense, just as you can ask a person for help and review.

But not using it for creating lies and pretending you're skilled in areas where you're not.

Or would you say that if HR uses humans to screen CVs, you can cheat by using a friend's CV instead (using a human, like HR)

>How is it not logically consistent?

i used my words to speak to the candidate, so they think its fair game to use their words to lie.

screening using AI could be a totally legitimate usage of AI depending on how its done. cheating/lying has no chance of being legitimate. just like speaking can potentially be used to lie.

most people here arent straight up vilifying the use of AI, just certain uses of it.

That's missing the point. The goal is to have a level playing field for the interview.

If your interview format allows people to use outside help but only if they think to ask, that's hardly a level playing field. You're testing the candidate's willingness to ask. In most interview formats it would not be acceptable to Google the answer, so most people won't ask.

If you have an interview format that allows Googling, you should mention that at the start. Not leave it as a secret for people to discover.

Every big tech company in the US has their version of PIP, to protect them from potential law suits as I understand, that lead to a situation that low performers might occupy their positions for months and years.

I think there's a tacit expectation to fit into a mold and that mold is heavily skewed towards extroverted neurotypical traits.

Seems pretty alien to my experience. A lot of senior talent was accustomed to traveling a lot anyway. I’ve certainly always interviewed in person and would probably turn down any company that didn’t offer as an option aside from COVID. But maybe there were a lot of companies that were willing to compromise on face to face so they could get any supposed talent to sign on the dotted line. Of course, they didn’t have much choice for a time even if they subsequently laid people off and/or largely froze hiring.

Do OCR systems still detect invisible words? I would have thought by now they'd use pixel based image recognition.

Interesting, that still works? I first heard about that a decade ago.

> would flag you internally

And how would they figure that out if you lie by exaggerating your experience and skills and not outright making up entirely false stuff?

Even reneging on an offer in the US gets you blacklisted for like 5 years max. It's not personal, it's just business.

The problem is that if everybody lies and you're the one not lying, you're worse off. In that scenario, the choice is between lying and being on even footing with everybody else, versus staying honest and getting an unfair disadvantage for it.

If enough participants lie, some of the honest participants get pushed out of the system, which makes lying more socially acceptable, which causes even more participants to lie... and so the feedback loop goes.

So what would be your advise to a fresh graduate (or even an experienced person) whose resume says experience in ".NET 3.0" where as the job posting says experience needed in ".NET 3.1" ? Remember it's HR or some automated system that does the screening.

Well if the idea is that the lying (to a sane degree) is only necessary to pass the “filter” and that it has limited impact on the candidate’s ability to perform the actual work its not necessarily that straightforward.

I mean, we have companies out there posting "entry level" positions and demanding 10 years experience in a technology that's only existed for five.

All bets are off, man.

AI will often casually lie / make up points which sound authentic.

> what is one to do

Find roles where your skills match the required skills ?

Recruiters have gone beyond keyword matching, they're now using AI to judge the resume.

> Hiring is broken

How would you, as an employer, filter out the frauds?

How do you reconcile that opinion with the fact that LLMs trained on programming concepts generally give incorrect answers about 50% of the time?

Is it actually more useful than Googling, or is it just so convenient that you let it convince you that it was useful? Or, depressingly, is Google just becoming so useless that something wrong a solid half of the time is still better?

> if you want to prepare for a C dev interview

Spend an hour reading a book about C?

I have a young colleague who wanted a job at a FAANG company, and asked for advice. I said spend a couple weeks studying the leetcode books - it will be the best value for time spent you'll ever get.

He did, and got a $300,000 offer.

Or you can open any good C book and review that way. Not to bash on the use of AI, but there's a lot of alternative ways that for me is more reliable to get knowledge from.

I mean, as an alternative to 133tcode, perhaps they could maybe demonstrate real world applicable skills:

* ability to communicate

* ability to empathize

* ability to be a nice person you’d want to see every day

Many interview coding questions are purposefully worded weird with the intent of seeing if you ask clarifying questions.

The interviewer might be looking to see how you deal with bad specifications which, in my experience, are also often confusingly worded, vague and/or conflicting.

Because interviewers don't care. I have tons of code on my personal github page and even thought it's listed on my resume/linkedin, no interviewer has ever looked at it. This includes contributions to a widely used OSS project.

YMMV, but all the high paying jobs I've received were due to knowing the tech stack they used and being able to walk through the projects that I've done in detail.

Admittedly, the last time I changed jobs was 2024, so things might be different now.

Everything I've written in the past 2 years is either proprietary or deeply private to me.

Me, too.

But, sadly, OP is right.

When doing a technical screen I'll sometimes pick a skill the person claims to have, and ask them the simplest possible non-trivial question I can ask.

For example, let's say you list 'SQL' as one of the skills on your CV. I might show you a SQL statement like:

  SELECT id, start_date FROM employees;

I'll tell you id is an auto-increment field, and ask whether the result would show the newest employee at the top or the bottom.

You have a 50/50 chance of getting it right. If you get it wrong, I'll tell you the answer. Getting it wrong wouldn't disqualify you.

Then I'll ask you how to get it in the opposite order.

I am expecting you to immediately say 'add DESC'. If you can't answer that question in under 2 seconds, you probably haven't written enough SQL to justify listing it as a skill on your CV.

You would be surprised at how many people fail simple tests just like this one.

(I won't use this particular one again.)

Yes me too, zero, and I'm pretty sure it's closer to zero than to 100% among others also, here where I live

I'm together with sethammons in that "rounding error". I actually go further and explicitly list things which I'm not good at on my resumé.

Thanks, interesting point and it hadn’t occurred to me that it would read like a direct quote.

I can’t edit it now, so will leave this here to say that it’s not a direct quote.

Well, let's see you do it! Can you deblur something like the first 2-3 letters of a name?

> Definition of being smart is to be quick at mental math and logic

That's not smart. That's being quick at mental math and logic.

Very different things

Yes, but we had that problem before when somebody would farm out coding assignments to a friend. I couldn’t say yet how it’s impacted the coding assignment’s effectiveness as a filter yet. We still do get crap code just sometimes it’s obviously AI generated.

That was not clear from your original comment. I read it as the company gave them the questions to presumably think about/prep.

That's all fun and games until a single company puts the top 3 or 5 candidates pitted against each other to see who waits the longest without a rejection and takes the lowest offer...

I heard this from friends, and despite being very comfortable where I am, I started interviewing cynically with no intention to take any job. I can confirm this is very much true and widespread. Hiring is at its worst ever.

Whenever supply and demand gets fixed, we'll see these behaviors go away.

Completely agree with the distinction you're making: framing is fine, fabrication is a deal-breaker. It's frustrating how often people conflate "putting your best foot forward" with just making stuff up, especially when they underestimate how easily it can fall apart during reference checks or follow-ups.

> Lying isn't marketing. If you lie in marketing, people can sue you.

That's why we had our guys down in marketing come up with a new term for it. Focus groups, legal review, the works! Now we call it "puffery".

> Lying isn't marketing. If you lie in marketing, people can sue you.

That is also fungible as well. Some lies just aren't catchable, like experience with skills that you teach yourself quickly, or go through a quick online course. Not saying I should, but "fake it till ya make it" is a definite thing.

> If you don't have references who can speak to the work you did, I also won't be hiring you.

There's also a reason I'm leaving the role, and usually you don't want people near your position to know youre looking.

And also, demanding references is the old AI slop - you're only going to give glowing references. Nobody gives bad references. And the worst case is you have a friend answer, or you buy one of those reference services (yes, theres a service for that).

> know that is frustrating, and I am sorry you are in that place, but lying will not help.

I think you're missing the point of the type of 'lying' I was referring to. Workday uses an absolute terrible AI, that uses keyword search. With my resume, the human readable text is accurate and me, but to this ai-slop scanning woukd scan 1pt listicle of every keyword.

Its not lying, but it is. Play stupid AI bullshit games, get gamified AI slop solutions. And I hate it. But even having a discussion with someone would be a start.

Leaving short-duration jobs off is common practice. The only way it might be "lying" is if you happened to, I dunno, have joined SVB just in time to commit a bunch of fraud, and then hoped no one googles your name. And even then, if it was three weeks, when your conviction comes up in the google search no one is going to think you lied leaving it off.

Similarly, it is typical that people will have a polite fiction for "why did you leave your last role?" that hints in the direction of the real reason without saying anything the company wouldn't want to be said publicly. That question is a test of your discretion as much as it is making sure the same reason doesn't apply to the new company.

However, saying you have a degree you don't, worked on a project you didn't, implemented something you didn't, led a project you only participated in, or used a technology you didn't: those are lies. Even if you get away with it, you are setting yourself up for a role you are unqualified to have. If you get caught, you will be correctly fired.

The pagination was chosen because it was prominently featured in the candidate's resume, and was something that interviewer was familiar with - it's not about "cool experience", but rather conversation starter.

When performing interview, asking about things mentioned on the resume is a pretty good conversation starter. No one wants random trivia, resume entries, especially from the most recent jobs, are absolutely fair game. And if they turn out too simple, we can always dig further based later.

What’s the better way to do pagination?

Well I guess if the candidate would be a little be stronger and actually trying to reason with the LLM about the decision it suggested, he would be better prepared and maybe got away with his claims.

Or as current best chess player Magnus Carlson said, "if I would cheat, you would never know". Meaning very strong candidates will get away with flexing the truth with AI. But this means maybe, you shouldn't look for a perfect fit. Or check his merit by spending time and money to get in touch with his old companies.