If you want me to spend a few hours performing a coding exercise, I'd be happy to, as long as I first get a few hours of interview time to ask you questions about your company and determine if it's a place I'd like to work. You're, of course, welcome to ask me questions about myself during that interview time as well.
The exercise is valuable to us in our decision making process, which is why we request that candidates do it.
It's disappointing to hear that basic professionalism isn't as common as it should be - but don't blame the exercise for that. Perhaps instead, breathe a sigh of relief that you won't have to work with people who do not value your time.
Yes, blame the exercise for that, because the exercise is fundamentally and inescapably is an attempt to assert a level of power over candidates; instead of studying their Github or just talking to them hiring managers demand that they take on the time-and-effort risk of qualifying themselves to the hirer. Which is really stupid in a seller's market, BTW. But, even past that stupidity, it makes it tenable, in the mind of the evaluator, to blow off the candidate because they have, by accepting it, signaled that they will put up with at least a decent amount of shit coming down the pipe towards them without serious complaint.
It's fundamentally dehumanizing hoop-jumping, which creates predictable results, and we should be better than that.
I've never read into it the way that you and many others are. I consider it as a means of quantifying ability. You don't have time pressure, you don't have an interviewer standing over your shoulder whiteboarding with you.
I view those as positives - because while I can do those things fine in day to day work, I do horribly at them an interview.
That said it's clear that our experiences with this have been very different.
* And o
I'm good at these tests. That doesn't mean that doing them is economically wise.
I have had a few people refuse to do interview questions in my life. Both times they were principal level devs who knew execs at the companies. Both times they failed after a few wasted years. That's not true of all of them, but I'm a principal dev, and I mostly work on architecture and planning, but I don't hire people to my role if they can't code.
Maybe it's because the whole thing is nonsensical power plays on the part of hiring managers when candidates don't realize that it's a seller's market and they don't have to play the game.
That's not strange at all. It's much easier to "fire" a consultant than an employee.
Rather, it's a recognition that we are business partners and the relationship must be collaborative and in good faith, with a much different dynamic than employer/employee.
- The person you've hired has made friends. You might like them personally. Firing them will have repercussions for your team's morale.
- Onboarding is a good bit more hassle than interviewing. Just like how it's cheaper to find bugs in dev rather than in prod, filtering during phone interviews is cheaper than in on-site interviews is cheaper than post hiring.
- Constantly hiring people, discovering they are incompetent and then firing them has implications for your own ability to lead and develop a team. So maybe now I look like the type of team lead who codes fine, but just isn't cut out to deal with the people side of things.
- In a small team environment, every team member has to be able to move the needle for the project individually. If I only have 3 engineers and 1 is bad, can I afford to pair half of my good engineers with the one who is struggling?
I'd much prefer to work at a company with a strong filter out front that then invests in the people it hires. Much better to say "Gee, Joe is mostly a wonderful employee but not great with skillX, let's see if we can pair with a Sr to work on that during the next project cycle" than have constant churn.
Its not always possible to hire stellar teams (by definition half of all teams are below average). Still work can get done. Look at Pair Programming, look at Agile. Efforts to create a process where predictable work gets done regardless.
But then you leave anyway, because leaving is how you get paid, and the company is out those resources.
Unless an employer literally always pay top-of-market, there is a severe disincentive for companies to invest in you. (A non-trivial part of why I went into consulting--my rates allow me to invest in myself.)
I think you think I have a problem with interviews; I don't. I have a problem with interviews that require the candidate to pay all the freight. Candidates have less time and fewer resources than an employer. Hitting them with some stupid code exercise that they can never reuse just to be considered for inclusion into your august freaking house is disrespectful and awful, I think.
Level the playing field. It's a seller's market (tech workers are still very much in demand).
i don't remember the last time i had to write bubble sort or a binary tree by hand. but i do remember when i had to deal with salesforce api errors and how to deal with them. or even, how to deal with rabbitmq retries compared to thrift methods.
Trust but verify. I'm sure some companies overdo it but if you're going to be touching code in any capacity I want to do the best I can to verify your abilities in a reasonable amount of time. A 1-hour coding exercise fits that criteria pretty well.
I know several leads who can write great code, but they can't lead a project to completion for the life of them. And if you discover that during the interview, you have to determine whether it was genuinely their lack of competence or because of poor management. Personally, I'd rather spend that hour allotted to coding digging more deeply into their non-coding skills.
It's crucial for first level leads to be able to code.
The role name may vary between company, but I believe we can agree that there are different roles to fill.
That's exactly the kind of "senior" person we _don't_ want.
Where my opinion differs from yours manifests itself in industry in the practice of separating "technical leadership" from "people leadership." Your architects (and leads) should be people managers. They should have authority in how their subordinates' time is spent. They should be able to resolve disagreements, both technical and interpersonal, among engineers. A "technical leader" designing large systems/platforms to be built by a team that's led by a "people manager" is pointless. I mean, who wants to design something without the authority to execute that design? That's a toothless fluff position.
It would depend on the coding exercise; you cannot just have a 30min chat about a relevant practical example? I'm not sure I have met someone who can bullshit his/her way verbally out of a random practical coding issue which I can just describe in a few minutes and they can talk with me about. But that's not some 'pen down a tree balancing algo with the following time/space constraints'. Which you might not have been talking about anyway but which the OP was about.