Live data from Hacker News

Follow-up to “The dystopian world of software engineering interviews”

jarednelsen.dev

181–190 of 538 posts

Re: Follow-up to “The dystopian world of software engineering interviews”

#181
post #144

Earlier quoted context omitted.

Sounds like the problem is low trust. It's a marketplace (for labour) where no transactions are taking place, because the buyers and sellers are afraid to come out to the town square. What do your counterparts at the company think the source of their hiring difficulties is? If not flawed process?

It's almost always attributed to the lack of quality candidates. But when you reach the final stage with 3 people who you know can pretty much do the job, who you know have successful careers working in successful companies, and you drop them because they didn't exactly behave the way you would prefer them to in a high stress scenario, and keep looking, I think you effectively forfeit the right to claim that's your p…

So much this. Every place I've interviewed at wants cookie-cutter candidates.

I didn't think of the solution you had in mind, because I've identified another problem that you didn't consider. And somehow that takes me out of the running. Good luck with that.

Not to mention the nebulous and vaguely discriminatory "cultural fit" problem. So, even if I manage to ace the technical, I still fail.

Re: Follow-up to “The dystopian world of software engineering interviews”

#182
post #179

Earlier quoted context omitted.

As mentioned elsewhere, this is discriminatory against people with families and and commitments that prevent them from spending hundreds of hours in prep. Not to mention it affirms Goodhart’s Law, where a single metric- the ability to answer DS&A questions- overrules qualified applicants from becoming hired. Not to mention such interview styles can be gamed. Suppose a Flatiron bootcamp for DS&A questions becomes big…

> this is discriminatory against people with families and and commitments that prevent them from spending hundreds of hours in prep Maybe but its not the purpose to select people with families or commitment. You choose to have families or commitment, you have to deal with the trade off. >An arms race for more and more difficult weeder questions? Its always be an arm race. Why to expect otherwise ?

Because hiring doesn’t have to be this adversarial process. And work doesn’t have to be this dehumanizing race to the bottom that excludes qualified people who are being excluded by bad metrics.

Re: Follow-up to “The dystopian world of software engineering interviews”

#183
post #66

What is tragic here is the amount of effort that is now expended by young computer scientists into the game of algorithms. Whereas a young programmer in the past may have spent their extra days writing games in Basic or static web pages, hackers these days seem to do Leetcode until the wheels come off. We are failing to bring certain softwares into existence by effectively requiring programmers to their effort into t…

> What is tragic here is the amount of effort that is now expended by young computer scientists into the game of algorithms.

This is a funny topic to me. A decade ago, before Leetcode, Hackerrank, and Cracking The Coding Interview, it was common for developers to complain that software developers were losing touch with algorithmic knowledge. HN-type websites were full of discussions about clueless software engineers doing damage to companies by implementing O(n^2) algorithms that worked fine on small datasets but failed in production. It was popular to complain that modern libraries, frameworks, and programming languages were making developers too soft to write good code.

The industry responded by re-emphasizing the value of CS fundamentals and algorithmic knowledge. The training industry responded with easy tools to teach and learn these algorithms. The internet is full of easily accessible materials to teach these principles to anyone motivated enough to Google it.

Now, the popular narrative has flipped. Internet comment sections want to hate algorithms and suggest that programmers don't need to understand the basics. Just rely on the frameworks and libraries to do the right thing. Just Google the solution and weave the libraries together to do what you want.

> Whereas a young programmer in the past may have spent their extra days writing games in Basic or static web pages, hackers these days seem to do Leetcode until the wheels come off.

Young programmers don't wake up in the morning and choose between writing static web pages or grinding leetcode. There are more opportunities than ever before to learn whatever you want.

Have you actually worked with students lately? Leetcode style systems are a lot of fun for people. It's a straight to the point system that teaches algorithms and other fundamentals in self-contained brain teasers that are straight to the point. And it's trivially easy to Google for supporting training material. In my experience, the same people who are self-motivated enough to build web pages and games are frequently also interested in Leetcode style brain teasers.

People aren't choosing between one or the other. That's a false dichotomy. All of the highly driven students I've worked with have a range of experience from toy projects to Leetcode style learning challenges.

Re: Follow-up to “The dystopian world of software engineering interviews”

#184

I have always hated software engineering interviews. This is a popular topic on this site, but nothing ever changes. Not sure as I would call them dystopian, I generally thought they were a chance for a senior dev to parade his ego. I could tell stories... (As I'm sure everybody else...) I used to take interviews and I always knew the outcome. "Sorry you are self taught, we only want people with CS degrees." "Or you…

Btw you have a missing table error at the page: http://www.strategic-options.com/trade/portfolio/hardline

(this is the Portfolio button URL in the nav bar)

Re: Follow-up to “The dystopian world of software engineering interviews”

#185

I've personally seen no correlation between people who can solve problems and program when watched by another person versus people who can solve problems and program when left alone with a real problem, and ample of time to noodle on and resolve it on their own. The same is true for writers: most are likely not to be interested and/or able to do their best work when watched over the shoulder by another person. The /r…

> I've personally seen no correlation between people who can solve problems and program when watched by another person versus people who can solve problems and program when left alone with a real problem, and ample of time to noodle on and resolve it on their own. At most of my companies, we tried moving exclusively to take-home interview problems for this reason. Short problems that could be solved in 2-4 hours of t…

If in typical work settings people work in groups to solve work problems, then in typical interview setting interviewer should also work in group with interviewee in order to have relevant estimate of interviewee performance?

Such collaborative interview setup also addresses interviewee concern that interviewer may not value interviewee time.

Re: Follow-up to “The dystopian world of software engineering interviews”

#186
HOW IN THE WORLD is this guy unemployed?!

(A: Well, he won't be much longer!)

He writes extremely well. He summarizes a great deal of feedback and knowledge really well. I don't care if he can code -- though evidently he can. Who would pass him up??

Granted, there's the question of whether he fits any particular job opening. But still, there must be many job offerings out there that he could fill if only the people doing the hiring could see that he's got something very special going on.

If you work in a team, in a large company, you need great communications skills almost more than anything else. He's got that down pat.

Jared Nelsen is going places. Ten years from now, when I see his name in the news, I'll enjoy knowing that I saw this a decade earlier.

Re: Follow-up to “The dystopian world of software engineering interviews”

#187

I don't know, hiring seems inherently broken. Not just the algorithmic part, even the part leading up to it. You have to impress a recruiter or an HR drone by listing all the buzzwords that are currently trendy. Then, even after the algorithmic questions and interviews where you prove you can DO the job, you get rejected. Every company I've worked for had trouble hiring. I'm working for a company that also has troubl…

I worked at a company where the interview process was basically "ask them about their experience, successes, and failures, and do whatever technical screen you feel works." Other than asking about failures (not everyone does) it was just "eh, wing it, tell us what you think." The hiring manager had to then weigh the responses and convince his/her manager if it was a go (we once rejected someone on ethical grounds after they passed all the interviews).

And the funny thing is, that got us a pretty good set of people. Even funnier: both the very best of them, and the absolute worst of them, eventually ended up at Google.

I guess if you can spend what Google spends; have a ruthless hiring process; can afford to carry dead weight indefinitely; have a standardized engineering culture; and have an endless supply of projects covering every conceivable skill level and impact to the company -- then you can probably hire your way to whatever FAANG-like cohort you think will work.

But if you have those things you're probably already a FAANG-like company.

For everybody else, I am not convinced there is a better system than "everybody use your best judgement, and have people who give a shit do the hiring."

Re: Follow-up to “The dystopian world of software engineering interviews”

#188
post #4

Glad to hear he hasn't given up. I agree with his basic thesis that much of software engineering hring is a cargo-cult shitshow. But I haven't given up hope and have worked on developing a hiring process that is human, efficient, and effective. I've outlined it here before. Been using it and tweaking it for the last 3 or 4 years. I keep bring it up because it's a tested alternative to the usual death marches and dump…

How do you design your coding challenges? Can you give examples?

Sure. We start by noting in our job description that there will be a coding challenge -- on a laptop, not a whiteboard.

Most of our work is on web applications, primarily Rails. So, for that role, the challenge is to fix an issue in an implementation of Fizzbuzz on our sandbox Rails app. We give the candidate about an hour to work on it. It's not as trivial as the common Fizzbuzz exercise but it's a pretty representative sample of the kind of work we do.

As part of the phone screening, we check to make sure the candidate understands there will be a code challenge involving Rails and that they're cool with that.

There's a handout we present before we turn over the laptop that explains the issue and clearly defines requirements (e.g. create a branch with this name). We read a brief intro section then give the candidate time to review the requirements on their own and ask them if there are any questions. Candidates are reminded to read the README included with the project and encouraged to use the browser to Google stuff.

One tweak we made was to add check-in stages to the challenge to make sure candidate doesn't get stuck for a half-hour trying to open up the text editor or something. Every 15-20 minutes, a team member will check in with candidate and if they haven't hit a certain milestone (e.g. has started local server or had reproduced issue), we'll assist them in getting to it. At the second or third check-in, it usually turns into more of a pair programming exercise and at the end we'll review the solution with candidate so they understand why the problem is relevant to the job.

There's not a definitive pass or fail for the challenge. The candidate's performance gets scored on a simple 1-5 scale alongside the other competencies we're evaluating as part of the interview. We're pretty thorough about preparing the code challenge and consider it a good example of our "onboarding begins with hiring" principle.

Re: Follow-up to “The dystopian world of software engineering interviews”

#189
post #179

Earlier quoted context omitted.

As mentioned elsewhere, this is discriminatory against people with families and and commitments that prevent them from spending hundreds of hours in prep. Not to mention it affirms Goodhart’s Law, where a single metric- the ability to answer DS&A questions- overrules qualified applicants from becoming hired. Not to mention such interview styles can be gamed. Suppose a Flatiron bootcamp for DS&A questions becomes big…

> this is discriminatory against people with families and and commitments that prevent them from spending hundreds of hours in prep Maybe but its not the purpose to select people with families or commitment. You choose to have families or commitment, you have to deal with the trade off. >An arms race for more and more difficult weeder questions? Its always be an arm race. Why to expect otherwise ?

At a certain scale, intent stops mattering. You need to take responsibility for the incentives in the systems you create. To do otherwise is just negligence.

Families are kind of important. They're usually a large part of the reason people have a job to begin with. To callously disregard the impact of a system on families is exceptionally appalling. You should not do that.

Re: Follow-up to “The dystopian world of software engineering interviews”

#190

There is no other sane and effective way to filter the insane amount of FAANG applicants, only DS&A. Other non FAANG/Unicorns that don’t pay as well as FAANG can drop the practice of asking DS&A questions. But for those companies that pay really really well, there will be insane amount of people that are willing to get a job there, and DS&A is a “May the strongest win” filter (although luck play a very strong role as…

This, yes exactly this is the reason I am preparing. I hate it but whats the alternative?
Post reply on HN