> The highest paying software companies employ LeetCode style interview

Ugh yeah it’s ridiculous. We can’t hire anyone! Have you considered not putting candidates through months of studying to memorize a bunch of algorithm tricks and then grilling them for 5-10 hours total on those? No - that’s not what the cargo cult manual says.

I dunno if it's that ridiculous. Having to spend a couple of months studying to get a 100k raise, seems not a completely unfair trade.

You'll be working with other people who jumped over some arbitrary hurdle.

What kind of team is that? Are you going to be surrounded by mentors? Free of bozos? In a world of competent executioners?

Unlikely. Look at how terrible a lot of the products are that these companies make.

They're fundamentally selecting for the wrong set of attributes. Look at the turnover rate - how many projects go nowhere and collapse, the quality of the output. Nobody ever leaves a job, they leave a team. The evidence is all around us that this process produces total shit results but it's become too much of a cult to change it.

Whenever I'm looking around for potentially better work and I see companies doing that, I don't even bother. It doesn't lead to better teams, more competent execution or better products.

Why on earth would I squander my time with some random ship of people who know how to do a sliding window? I'm sure they're all friendly, but we're doing hard things here and need the right mix.

I mean what on earth, fix the fucking process and maybe they'll stop incinerating dumptrucks of money on badly executed technology

I don't know if FAANGs are so deserving of such vitriol. I would probably say that big companies overall just fundamentally have a set of issues, and FAANGs are probably above average in many aspects relative to the rest of the big ones. Now, if you prefer certain small companies, or other humanitarian or mission-focused ones, that's not a FAANG-specific complaint, in my mind.

And just product-wise, it's just way too easy to find counterexamples, like, how worse would the world be off if there was never invented google, gmail, chrome, iphones. Before google, there was no reliable way to search the web. Before gmail, there was no such thing as a dynamic web application. Before chrome, we were stuck with microsoft making up their own standards with IE and an inconsistent web spec. Before iphone, we were stuck with feature phones. The list just goes on and on.

> google, gmail, chrome, iphones

There's four excellent examples of things that were better 10 years ago before this practice was instituted. The youngest thing on that list (chrome) is 14 years old.

Both Apple and Google used to not hire in this 3-4 interviews of canned programming question way. Then they did. Those products are now stagnating and deteriorating so much that it's obvious to everyone.

If you think it's exclusively a FAANG practice, you're mistaken. There's tiny firms that do it and really large ones that don't.

Bringing the worst practices of standardized testing into the tech hiring process produces the same kinds of problems that standardized testing does elsewhere. Silicon Valley used to be better than this.

It's a dumb arbitrary ceremony - might as well be a physical challenge from Nickelodeon's Double Dare.

(Also your narrative history is wrong, but that's not the point)

> There's four excellent examples of things that were better 10 years ago before this practice was instituted.

FWIW, I interviewed with Google in 2009 and the interview was 100% pure leetcode (not even design/architecture questions). It was exactly a series of four back-to-back 1-on-1 meetings, which consisted of "Hello" followed by an algorithmic puzzle.

If anything they seem to have broadened the scope of interviews since then.

Alright. I stand corrected. I also went to San Francisco job interviews around 2005 or so and found that. I was turned off by them back then as well.

I'm not a researcher in the field but I'd love to see some study. Let's say measuring for deadline slippage, budget overshot, and retention rate.

The testing groups will be teams that use these tactics almost exclusively and those that use a variety of other methods.

If my assumption that this is a mostly arbitrary attribute is correct, their averages would be nearly the same.

If they produce statistically better results then I'd literally quit my job and go work for one. I'd be happy to be wrong about this but I don't think I am

I’m interviewed in 2006-2009 at all the top tech companies and can confirm the process was essentially the same as today.

Why are those 4 thing worse no?

Ok, I understand that google search is worse, but it isn't obvious to me that's a result of engineering vs executive choice.

I doubt the stagnation is the fault of poor hiring practices.

More likely has to do with monopolization and business models that favor advertising over users.

Can you elaborate? I'd like to hear your thoughts

Monopolies aren't motivated to innovate. (At least not in helpful ways.)

This is why the web is stagnating with near Chrome monopoly, just like during (shudder) the IE years.

Same for Google itself. Bing/DDG are only half-assed competitors. Google seems to be making search worse for all kinds of bad reasons, just because they can.

I'd guess most people who use an alternative to gmail specifically don't want to use Google, not because something else is better.

Spending a couple of months for a guaranteed 100k raise would be great. Spending those months for a tiny chance of that raise would not.

I thought the idea was that if you didn't spend months studying, you'd only have a tiny chance, regardless of how skilled you were. And that the months of studying is what's then required on top to get that offer from at least one of the big five you apply for.

That I commonly see complaints about how other people "without" skill were getting job offers just because they put in the months studying, seems to indicate that it's actually quite a generous chance.

It’s selecting for certain attributes that don’t have a major effect on job performance. I would actually paint it as pseudo scientific at this point.

Leetcode questions are not really selecting for anything. It's mostly selecting against false positives.

Yup, don’t regret spending 2 months for a $250k raise one bit.

The vast majority of tech companies including startups are asking algorithm style questions these days. It’s adapt or die.

That's the first interview. They usually test your ability to design systems and stuff in a second interview.

If you get it.

>* The highest paying software companies employ LeetCode style interview. That is off-putting to a lot of people even though the money would be super sweet to have.

Not only off-putting but out of reach to older engineers with family responsibilities. I would also add that the value proposition for working at a FAANG isn't nearly as competitive if based outside the US.

I agree. I would go as far as saying that LeetCode interview is biased against older engineers with families. Especially young families.

What?! Age discrimination in the software industry??? Say it isn’t so!!! Seriously though, your insight is astonishingly spot on: LeetCode is a perfectly legal means of practicing age/lifestyle discrimination. Yet another innovative Silicon Valley dark pattern. Why is this the first I’m hearing it characterized this way?

That's just it, it's stuff you learn in computer science, then forget about because you don't use it in your day job.

Your day job is writing CRUD services. That's the job.

Our tech interview - ~8 hour take-home assignment, it was well-paid consultancy - asked people to basically demonstrate they could write a CRUD service. It had reading XML, some business logic / calculations, a REST/JSON API and a front-end.

Ten years later, I'm still working with that basic model. I mean sure, it's been broken up here and there by a different tech stack (Angular instead of React, Go instead of Java, native iOS instead of web) but the basics are still the same.

I haven't had to invert a graph or do a depth / breadth first search since college. The most algorithmic thing I do is implement a comparator function for sorting.

Hopefully you mean the 8-hour assignment was paid or that would be an unacceptable ask.

8 hour take-home ? Shesh that's terrible.

Yeah, this is so obvious once it's stated this way, but I'd never thought of it like that either. Every time I talk to prospective employers and they wax about their rapid fire technical challenges and 5-round interview processes that last six weeks, I wonder "how the hell do they expect someone with other commitments to do all of this?" And the answer is ... they don't!

They ask questions that kids are taught at 18/19. It's almost shocking; this is how you're selecting?

People that are in their 40s aren't confused by the question but are more confused by why it's the question as being able to answer it is so disconnected from any metric of competency or quality.

It'd be like saying "so you have an engineering degree, oh a masters from 20 years ago. Cool, remember Lagrange multipliers and Hamiltonians? I'm sure you do. Here's a problem. You have 10 minutes"

That's why we keep seeing the failures of the past repeat. Tech selects for people too young to know it and too arrogant to study it.

I feel like by the time you’re in your 40s, you should be able to leverage your network to bypass the technical interview. What are you doing at 40 going through the same hiring funnel as 18 year olds with no discernible reputation or experience? Maybe if you’re doing a career switch, sure, but if you’ve got 15-20 years under your belt then that’s a different story.

You are proposing substituting a nepotistic process as a substitute for a discriminatory process? Personally, I do not think that would yield any improvement.

No, I’m just saying at some point your reputation should precede you, otherwise what have you been doing for 20 years?

Not only do some companies strictly have the policy regardless but also some people (like me) want a consistent challenge as in going from say security research to hardware to web dev.

I don't advocate for a free ride nor am I saying the tests are hard (they really aren't), just that they have a very weak correlation with building strong teams of the right people

Being an introvert? Becoming a better engineer? Learning new skills? Doing things that actually matter for my job?

Speaking as an introvert, introverts need networks too. Build a network - skip to the front of the line. Or better yet, get a job specifically made for you. Otherwise get into the funnel like everyone else.

Networking is a non technical skill. Hard to say how well it correlates.

It’s possible for already wealthy people though. Poor people need $x per month. A bit more wealth and cashflow is not an issue and so you can look at $x per decade. In that scenario taking a gap year to bash leetcode and interviews is perfectly rational. Its a 10% tax at worst.

So you think it better to discriminate according to the size of a candidate’s bank account? From where I sit, you are saying the poor do not deserve the same chance. I think age discrimination would be less offensive to the masses.

The opposite. I’m observing why the rich can get richer, because they don’t need to make ends meet. But I am not saying it’s a good state of affairs.

When I think of pros and cons of re-learning this crap I always re-evaluate what I have and what I want - and always come to the right conclusion - my time is better spent elsewhere - working on my own ideas or just more family time.

In the US, thats not generally true. It takes longer to prep from scratch if you have a young family. If you have an older family (and are even older yourself) that constraint is reduced. Realistically you can prep doing a few questions a week, it will just take longer. It’s not pleasant. But it’s more about the commitment, the plan, and enough research to justify it. If you can reasonably expect a 50%+ pay increase and that would translate into a better life, most people at most ages can make time. It’s about experience and practice with generally well known algorithms on well known problems. It’s not rocket science or intelligence nor nearly as much prep as many other high paying jobs (my MD took a decade of study for instance). Yes, annoying as hell. But not insurmountable.

* The people who end up winning your Leetcode interview will have the least amount of trouble leaving your company.

That's why if you're a small start up, trying to hire like Google is the dumbest thing you can do.

There's something much worse than Leetcode. Take home unpaid long projects that end up in rejection or lowball offers.

I had a take-home assignment not long ago for some... dodgy recruitment-consultancy-disguised-as-training company. It was the exact same assignment as my previous employer used ten years ago. Ended up just showing and talking about my current project, which is basically the same thing.

Definitely. My last job prospect would have required me to lock myself in the room for a weekend or more doing a chain of low dopamine tasks.

I have done take home projects before, which often resulted in low ball offers. At most it made me angry for the time wasted, but now having children I just don't even bother anymore.

Related to the last point: There is no way to estimate how much work a new job is in reality. I'm currently working 10-20 hours a week remotely. I think my salary could be improved, but taking a risk of actually having to put in 40 hours or more of work every week is not worth the small(ish) increase I could reasonably get by changing jobs.

People have been bitching about LC interviews for 10 years. That's eons in tech, yet companies continue to use it. Maybe, because it works extremely well? Tech companies are taking it in more every year.

Depends on how you define works. It does effectively filter the people that can't code, but it also filters a lot of people that can code and just can't or don't care to do it under time or environment pressure.

In the past that was fine, candidate pool was large enough that a heavy-handed filter was okay. That may be changing.

> Depends on how you define works. It does effectively filter the people that can't code, but it also filters a lot of people that can code and just can't or don't care to do it under time or environment pressure.

It is much more important for a company to filter out likely low performers than it is to avoid filtering out likely high performers, because the costs of hiring an incompetent person are enormous. Not only do they not produce as much, they eat up others' bandwidth by asking too many questions, implementing super buggy code, or not communicating effectively. They make other people on the team feel like they're not pulling their weight. They're a pain to fire depending on company culture and laws.

This process is designed to overfilter and maintain an A-player roster as much as possible, vs putting butts in seats. It's hard to tell whether the pool is exhausted here - by some indications, there's a lot of attrition but also a lot of hiring...

This is only true for companies that have a large talent pool e.g. FAANG companies. For some small 30 person business in the back of nowhere that wants you in the office 5 days a week they cannot be anywhere near as picky.

I tend to think it's even more true for smaller companies that do not have the margin for too many mistakes. FAANG do not have an all-A cast. They have enough buffer to weather the odd mishire; and they also have enough projects going on that they can find better projects for people based on their strengths. At a small company on the other hand, the wrong hire can be existentially problematic for the business.

Mistakes can be fixed if they are small scale i.e. one bad hire stand out at a small company and can be fired much more easily. What you're missing is that for these small companies they are excluding the best candidates with this process. In terms of bad hires at large companies they are harder to notice and large companies have more process that tends to make firing people harder.

This is why these big companies do it, because it is actually quite hard to be fired from someplace like google whereas at these small companies you'll be working directly with the owners.

Also people will do all these take home tests and 5 rounds of algo interviews and CCAT tests ... etc for large well known companies as it is prestigious to work for them. They literally won't do them for small companies. For small companies unknown companies all you'll get is desperate people you are actually filtering out all the decent candidates. I went through this process last year it's very hard to hire as a small unknown company.

Every job related thread on HN devolves into a bitch fest about interview questions. I don't know how it works at non-FAANGS, but Google gets literally millions of applications per year, and even the maligned FB gets a few hundred thousands. I have yet to hear any workable alternative hiring process that can cope with such volumes. At these companies, the hiring probability for most individuals must by necessity be very low, no matter what kind of interview/hiring process you'd suggest.

People have been bitching about a lot of things for a long time and they’re still in use… amount of bitching is just not correlated with anything.

More likely it works "good enough" rather than extremely well.

The companies that pay 80% of FAANG salaries without the LeetCode whiteboard bullshit are out there.

Can confirm, we do a low-pressure, non-timed, non-puzzle, somewhat realistic, short take-home exercise. But we don't (and can't) add 20+% to your salary in RSUs.

20% seems very low to me. In the 2 FAANG's I've worked at, the RSU's ended up being 100+% of my salary (which was average for my level).

Initial grant was 100% of one year's salary you mean right? So +25% of yearly salary before appreciation?

Otherwise you're saying something like a 250k base, 1MM RSUs before stock appreciation... which is a very lopsided looking offer

The non-faang but with faang pegged compensation company I work for does yearly grants

My L5 comp is ~$150k base and $140k equity per year (I'm based in UK, would be higher if I was in the US)

I guess this is just lower base than I'm expecting

I work at a company in a similar situation (we get a lot of Xooglers and pay accordingly) and my L6 base is 250k and I have 150k worth of RSUs vesting per year for the next 3 years.

I do negotiate base aggressively vs equity so that might be skewing my view of things

L6 (~ Staff eng) new hire offers at rich big co usually are af 250 base and 250-300k/yr in stock. See levels.fyi for some recent ones.

250-300k and 250k base still seems pretty uncommon for my role (mobile development)

The only company I see doing that somewhat consistently is Fa.. Meta, and they're slipping into the same place Amazon is, where they need to pile on TC to make up for an image problem

I can tell you this is not true.

It's common at top companies - they are all competitive with FAANG comp. You will easily pull in $400-500k in total compensation as a mobile dev at sayyyy an AirBnB or similar. Roughly even split between base and equity.

Isn't your comment confirming what I said?

If 400k (where I am with 250k base) to 500k (where I'd be with a 50:50 split) is the entire range you're mentioning, that's the same as saying "250k - 350k RSUs with 250k base is uncommon" isn't it?

-

It might be relative to my situation: I have 8 YOE experience, I'm in mobile, and most people I can talk to candidly about salary are similar. So the numbers I hear directly might be tinted by the fact ~8 YOE isn't usually enough for band topping L6 offers

That being said, even Levels.fyi isn't showing me mobile devs negotiating 250k base without giving up anything on RSUs is a common situation (it definitely wasn't when I got my job, but even with the current market craze that feels like a stretch)

If I'm missing something I'm glad to know more, one undersold upside of the current market is you don't absolutely have to job hop to push for more comp

Yeah I think that's fair, maybe I misread your comment. I find that base maxes at $250k but equity has a much higher ceiling as the eng levels get higher.

I meant 100% yearly, but after reading your comment I double-checked my offers and I was wrong. RSU grant (per year) was at ~45% of base salary.

I was going by my tax filings, and because of the stock growth (and additional grants), I end up with 100+% yearly, but that's quite different from the initial award.

Thanks for pointing this out.

This is actually a common offer. Some companies that wanna continue to offer high compensation without this upfront sticker cost are now resorting to one year rather than 4 year grants (stripe, Lyft).

Stripe and Lyft are definitely not aiming to avoid upfront sticker cost with their 1 year grant scheme.

An exceedingly charitable view is that they're limiting downside, but realistically they're limiting upside so that you don't end up with people reaching targeted compensation for the next 4 years in a year or two and then leaving regardless of unvested shares

-

But I also mentioned in a comment above, I think my view is skewed because I negotiate base aggressively

As an L6 I make 250k base, and my initial grant was just over 600k in RSUs at a non-FAANG company. I think that's a higher base than median but lower equity than median for FAANG

Why? Unless you’re sure that the stock price will tumble, it makes no sense to increase your base pay aggressively. Most higher engineering levels will top out their base in the 200-300 range and all the rest of compensation will be via RSU.

We're barely a year removed from some of the most volatile stock market action since 2008

Barely removed from a lot of people taking massive haircuts as COVID darlings are down 50%+ from ATH

-

To me equity is an investment you can't back out of for 4 years.

I'm not sure the bottom isn't going to fall off the market for tech period, so I'm happy to trade equity for the ability to have some control over my investment

It can be over 100% per year depending on level (more senior = more equity). Check out https://www.levels.fyi/

If you're senior L5, it can be around 100% per year even before appreciation. Above that and it starts going above 100% per year.

Great Senior Engineers are making 370-500k. Are you paying even 80% of the lowest end of that range?

Can you tell which company this is? I would like to know.

Do they typically offer RSUs? I'm asking because my RSU awards double the total compensation, so if you're counting 80% of just the salary part, then that number would end up being 40% in my case (and I assume for everyone else at my level).

My employer offers RSUs and is a public company.

Do you have any pointers for those who’d like to filter for that?

I recently came across this list which might be useful: https://github.com/poteto/hiring-without-whiteboards

It's really not that hard to prep for interviews.... Why take the paycut?

The prep itself may not be that hard. Depending on your non-work life, finding the time to do said prep may be difficult. I would like to double my comp. I also have young kids who occupy most of my wake-work and then work-sleep time. There are maybe three hours at the end of the day that aren't spoken for, but then there's housework, spouse time, and just being tired. It's difficult to fit "grind leetcode" in there. This is due to choices I've made, and I'm happy with those choices! I blame no one but myself for them. But I also think I'd make a good employee at leetcode interview companies were it not for the leetcode, and I think there are a good number of people like me.

If it’s helpful i got into the leetcode grind in a similar situation by getting up 60-90 minutes early and doing 1-2 questions a few times a week. Thoroughly understanding top 50-100 questions will take you very far. So long as you do similar questions in groups (ex: all binary tree questions) it makes it easier to understand and retain. Then even if you take long breaks between study (months), you’ll mostly retain the underlying techniques and can just refresh when it’s time to interview (ex: a week off work for a final review). it’s very annoying. But don’t sell youself short, you can mix the grind with breaks to retain some sanity. Just make sure to have a plan (ex: do top 100 questions )

Spoken like someone without kids or external responsibilities. Free time is a precious resource for those with families.

Looking at it internationally, I could earn 3x by FAANGing in the US, but I would be on a sponsored visa and massively disrupt my life if fired. The risk factors put me off.

Money should be a big factor. Imagine a job where year 1 you work and year 2,3 you get to do whatever you like and own 100% of the equity. That is what 3x TC basically could mean.

PIP culture? Google isn't giving me any answers here.

PIP stands for "Performance Improvement Plan". PIPs are a relatively common tool to give employees advance notice about getting fired, but Amazon has a notably brutal setup that uses stack ranking. Described here: https://www.reddit.com/r/csMajors/comments/kh9ztt/can_someon...

posting just as an acecdote, but as a AWS HM PIP is only used for really really obvious under performers. Honestly, one regret is not putting under performers on a plan fast enough because it is a big drag on team morale when someone isn't pulling their weight.

“Performance Improvement Plan”, it’s what you get put on when your company wants to fire you but wants to show due diligence that they tried to give you a chance so you can’t sue them. By “PIP Culture” I take it to mean Amazon where the expectation is some % of the team is always expected to be fired so PIPs are a frequent occurrence.

Picture in Picture is the most common use of the acronym, but here they mean performance improvement plan, which in many places is basically a kiss of death.

or pip install :)

Yeah - if a company does leetcode and cannot justify it, I'm out. One justification would be that the job actually requires understanding/writing a lot of algorithmic and very performance sensitive code, but that's not the case for the vast majority of jobs.

And I do the same for all other parts of an interview too. If I were asked to knit a scarf during the interview process, my reaction would be similar.

Can you define PIP, please?

Performance Improvement Plan https://www.perkbox.com/uk/resources/blog/a-complete-guide-t...

"Performance Improvement Plan". The "PIP culture" mentioned is a culture where there is an (perhaps unspoken) quota of people that need to be put on PIPs, so they are quick to pull the trigger. Very uncomfortable as it keeps people on edge.

Personal Improvement Plan. Usually used as either a blunt instrument to goad people into working harder or a sneaky way to fire people. Topped with a sauce of double-speak of course.