Hacker News Reader

Comment by Aeolun | Hacker News Reader

You think that’s unbelievable? I’ve literally had a company recruiter tell me that I should probably spend a few hours on leetcode and hackerrank before the interview.

That turned out to be predictive, because they rejected me mostly based on my inability to do well on those kinds of tests.

I got a life people. Between taking care of my son, wife and work, where am I supposed to find ‘a few hours’ to spend on something that is solely meant to do good on your company interview.

It could theoretically have been worth it because of increased compensation, but something in me just screams nooo whenever I consider it.

View on HN

You think that’s unbelievable?

Google and Facebook recruiters told everyone at my university they expected people to spend 2-3 months of several hours a day leetcode type practice before applying. This was said in a panel type "how to get a job at FAANG" seminar for graduating seniors, with actual Google/FB SWEs present, and all nodding in agreement that this is the norm.

That time investment sounds like a tragic loss of open source development and random curious exploration by students.

Maybe also quite a few lost innovative startups, which might've been spun out of students making things they want to make, rather than grinding just to prepare for interview rituals.

It can be a massive fun too. Ive never interviewed at FAANG, but solving some leetcode problems (mostly medium-hard though) was so cool for me and some of my friends. We were into olympic math at school, and I don't consider that a waste of time as well.

Good point. Practicing those exercises for sports-like competitions sounds great for people who want to do that for its own sake.

But there seems to be a large societal opportunity cost to everyone spending several months practicing doing that on a whiteboard, specifically for job interviews.

It's fun but for the most part boils down to software trivia, similar to Quiz Bowl. The key to succeeding with leetcode on the harder questions is memorization of things that frankly anyone can google up pretty quickly. It's a poor reflection of actual work.

Lol, you win ;)

That sounds terrifying.

I think it's definitely worth it for some people, at least financially. But many many others will study for months just to find out they can't make it no matter how much they prepare. I mean they accept how many - 2% of applicants? If you can look at it as a learning experience that's fine as well.

> If you can look at it as a learning experience that's fine as well.

A gambling addiction is also something you get from (in)appropriate learning experiences.. I think Nordvig's view of programming and materials that fit a longer view of learning good style are likely to work against getting a job in this climate. You are more or less demonstrating your willingness to embrace the anti-patterns a company holds dear to the point that you could be stuck with them without even being considered for the job.

A normal interview process probably costs 4-5 hours just to do the interviews. People often travel for them. You spend all day doing your job for years on end. A few hours of prep doesn’t sound like a big deal, especially since that prep translates to other companies as well.

Having a life means that you should love a standardized interview format. There's nothing worse than project assignments from companies asking 'for only a few hours' with descriptions that very clearly ask for multiple workdays.

Algorithmic interviews also compensate you better. Being able to prepare for multiple interviews at the same time means that you have a broader range of options to choose from and can play them off of each other.

Home assignments at least test your actual ability to write code or complete something that somehow resembles real work. Whereas solving whiteboard / live codility questions under stress with some guy looking at everything you type is a lot about how well you are under stress and how used you are to solving codility questions.

Perhaps. But from the employers perspective home assignments carry a much higher risk of cheating

I suspect with some startups is to do with the ecosystem. At first glance it makes no sense your tiny no-name startup adopts a FAANG interview process; it's like running a US Marine bootcamp for your local cub scouts. However the best realistic (non-unicorn) outcome for a VC-backed startup is a FAANG aquihire, and the acqusition process may depend on the pedigree of your engineers and their ability to pass a leetcode interview as part of some pre-acquisition technical audit. Things like being able to execute on good architecture and design and overall code quality - the mark of experienced developers as opposed to newly-minted CS grads - are of lesser importance, as the product itself will probably be quietly shelved in a year. In other words the FAANGs see the startup ecosystem as a cheap recruitment funnel, so it makes sense for founders to optimize their interviewing process around that.

> the acqusition process may depend on the pedigree of your engineers

I doubt it. Investors care about sales, growth etc. If they really care about the technology itself they'd probably look at the CTO / tech leads and thats it. I don't see how some developer's interviewing skills matters to the bottom line.

Reduce your daily TV watching and/or social media interaction by one hour, and you now have one hour a day to do this. Aim to get a remote job so that you save another hour or two from the daily commute. I did this. It works.

why do you expect to pass an interview without preparing for the interview?

Maybe in the hope that the interview would reflect the actual work that needs to be done, which my education and previous work experience have prepared me very well for.

Then I'm happy to prepare for the interview in the sense that I try to learn as much as possible about the company and the people I'm going to meet.

Copying this response from my older comment:

The rampup at FAANG can be brutal and you'd be doing lots of learning even after getting in. The leetcode style interviews are designed to select people capable of studying and materialising broad problems in depth and scope under time constraints. They are also objective and can't really be faked nor you can talk your way in like in most banking/finance IT jobs. I was of similar opinion at some point but these realisations made me change my mind and I've come to like leetcode interviews. But I also think they only make sense in FAANG scoped companies where all the stack is proprietary. Small companies or companies using opensource tools don't really have this rampup and are doing such interviews mostly because others (FAANG) do it.

They select people who have a lot of discretionary time. I guess that is ok, if you want an organisation consisting mostly of young people without family that are happy to work weekends and evenings.

I’m very happy to be in a position where I can do most learning during company time, and then spend my free time with my wife and kids.

I can definitely agree, IT industry sucks in the sense that you're sort of expected to always be learning to stay relevant in the job market, and on your own time at that. We don't expect dentists or lawyers to have their job be their hobby but in IT... Here we are.

The assumption that leetcode is the only way to know if people have learning capabilities is weird. My job is a semi niche industry (semiconductor manufacturing equipment), most people who enter here need to go through a brutal ramp up. There are ways to assess learning ability without asking people to make interview preparation their part time job.

Edit: Anecdote time. The best interview I ever had was for my current job. It was a hour long where they gave me 1 design problem to solve. It was an open problem but the interviewer kept introducing new information. "Lets say the speed of stage cannot exceed x", "what if the wafer cannot take the stresses you will generate?", etc. I think that showed both my design thinking and the ability to assimilate new information in my mental model.

FAANG typically has 3-6 30-50 minute back to back interviews, the style of interview where they ask open ended questions is almost always included too...the leet code stuff is in addition to that. I can agree it's not ideal but these big companies do it for a reason and apparently it works for them.

If design thinking can be assessed through open ended design questions, what is leetcode adding? Except proficiency in leetcode?

The selection is a filter for people who are willing to pour massive amounts of free time into preparing for a small subset of companies. It's a great filter if you want someone who obsessively wants to work for you and is willing to sacrifice a significant amount of their personal time for it, but not so great a filter of their competence beyond a basic level.

Leetcode is a memorization/trivia exercise, not a reflection of someone's deep understanding of proper coding practices, software architecture, and ability optimize in real world cases (since in the real world you can use google and use whatever libraries you want).

I think you are genuinely curious. The point here is, people are confident that they can do the job, just can't pass the interview. The purpose of the interview is to select someone to do the job, not to pass the interview. "Preparing for the interview" is chasing the wrong KPI for both parties. Companies don't want someone who can pass the interview, they want someone who can do the job, they are not the same things.

However confident people are hat they can do the job the point of the interview is to convince the recruiters.

That's the game and the recruiting company sets the rules.

Thats the point though. The recruiting company rules seem to be suboptimal and infact optimizing for something unrelated to that job.

Edit: Fun example. Its like recruiting companies are trying to hire for the next big chess tournament and their criteria is how fast someone can win practice end game scenarios. It might seem fine on the surface, but the truth is end game is only 10% of the game.

> to convince the recruiters.

The employer is the one with a skills need. Arguably, I'm in need of resources. However, coding shops are a dime a dozen. I know I can build and ship software. I've proven it to get where I am today. I'm not too interested in proving it to someone who will forget my name in a week. I've also proven I can lazily apply to places, ignore interviews with this coding challenge mess, and pay attention to the shops that actually talk to me about my strengths and their needs before anyone opens a CLI.

So, is there a STEM shortage? If yes, then the recruiters and companies are the ones hurt by the 'convince the recruiters' mindset. If there isn't, then I'm hurt by being less competitive. Who knows what the future will look like, but despite all the language and shenanigans to the contrary, I know my company needs me more than I need it.

The recruiting company sets the rules for their own interview process, and most developers have an abundance of companies to choose from.

Obviously, the whole complaint is that the rules they set are a poor indicator of future employee performance. Yeah, they can set a rule that whoever is the most physically attractive gets the job, but again, it's a very poor indicator of employee performance. Do you follow?

idk, seems to me weird to go to an interview without preparing for days/weeks for it.

It doesn't seem weird to you that the skills you need to prepare for the interview have nothing to do with what the job actually entails?

Some delta between these two is true of all professions. How is the employer supposed to know what your skills are? Assessing this is a skill and there are only badly flawed tools available.

Every time this kind of argument is made, I am tempted to remind people that all other professions do without this kind of thing (except finance I think). But then I remember only FAANG type companies require skills, bridges are built by randomly selected people.

And those fields will have plenty of capable people who don't get a shot with the best firms because they didn't go to the right schools, or know the right people, or use the right buzzwords on their resume.

The coding screen is a less dirty shirt in a pile of dirty laundry. At least it's an objective measurement, even if it doesn't measure precisely the right thing.

Reply on news.ycombinator.com