Nope. My wife just makes sure she knows where her physical portfolio is (basically just print-outs of old proposals she's worked on), spends a couple hours reading up about the company and maybe rehearsing what answers she might give to some common behavioral questions (the cliche ones, like "What's your biggest weakness" type crap), gets her clothes ready, and makes sure she gets enough sleep. This is the night before, usually.

I've never seen her spend more than a day or two preparing for an interview. And she gets paid near the top of her position, and has worked on dozens of multi-million dollar proposals.

The only reason it takes her a while to find a job is there's not a ton of positions that are willing to pay her what she's worth now (she's gone through job interview processes only to be given an offer half of the bottom of the range she provided multiple times), and she usually has to try to sell them on what she can provide to the team that they weren't even looking for (but should have been). Doesn't always work, obviously. Budgets aren't always there.

Meanwhile I've spend 2+ weeks or more preparing (hours each day) for SV interviews and still failed them. In Facebook's case, I was asked three questions by an HR person and gave the right answer to the third question, just not the term they must have had written down for them on the paper, was told I was wrong, and then got a rejection email. At Zynga I was asked to come up with and speak the code to a permutation algorithm, over the phone, on the spot, without using a computer. At Google, in one of the six interviews, I was asked how to convert a 3D grid of buildings into a 2D skyline with a single line, and I had difficulty figuring out the logic for overlapping buildings in less than 30 minutes using only a whiteboard.

I always spend at least 2-4 weeks refreshing algorithms and technical trivia questions before I even start looking for a job in tech.

OK I'm curious, this is computer graphics crap like z-buffers (I would probably say form a binary space partition tree, color all the polygons one color and a default background as black, rasterize the view, and then trace the outline)?

Or is this some trick?

I didn't figure it out in time and he didn't elaborate, just sat there and watched me flounder, so I don't know for sure. This was way back in 2012 too, so I don't remember the details of what he asked super well.

I was under the impression he was wanting me to figure out how to do it using only the bounds of the rectangles mathematically, and not by changing how it rendered, from what I can remember.

A 3d map of a city is a bunch of polygons. A skyline is a perspective on that 3d model, so you need occlusion/etc. The single line is then a postprocess on that view of the skyline.

Maybe there's some 2.5D cheats like DOOM did since the 3d model is all on one plane.

Agreed, a ridiculous question.

These questions where you either know the algorithms of giants, or are expected to reverse-engineer algorithms that people did research on for months or years... The height of arrogance.

What is your wife’s occupation?

Proposal/Marketing Manager, but she often does things outside of her official job description as well.

No. An interview process can be very selective without really requiring prep beyond studying up on the company somewhat--and maybe not really even that to the degree you're already familiar with the company.

Not a developer (though was an engineer) but I've never studied for an interview beyond maybe spending an evening reading up on a company. But, to be honest, the (few) jobs I've had over the past 20 years or so have always come through knowing people.

Some people use algorithms daily, or went to school for them, and don’t need to study for leetcode. Likewise, some people do architecture or frontend apps all day and don’t need to study for those. No matter what a company asks, some people will find it unreasonable, at the end of the day the charitable interpretation is that the company is choosing interview challenges that they feel are a proxy for the skills needed to succeed.

You're assuming that a company has to interview all candidates the same way regardless of what job they're actually interviewing for. A frequent complaint, voiced all over this thread, is that interviewing like that makes no sense. If the job uses 100-level algos all the time, then sure, let's talk algos (also, kinda weird, but I'm sure there's something out there like this). But if you're bringing me in as an SRE and we spend a bunch of time on LC questions and much less time on my experience with AWS, Terraform, and CI/CD work, the interview is bad for me (because I don't work with low-level algos and probably won't know that stuff, _unless_ I've been grinding LC) and bad for you (because you don't have much signal about whether I can do the SRE stuff).

I’m not assuming that.

I’m just suggesting some companies may want SREs who understand fundamental computer science and are able to apply it to novel problems, perhaps as a proxy for finding candidates who could create their own terraform or quickly learn whatever tool may supplant terraform in the future.

As an analogy, it’s valuable to hire an electrician who knows the whole electric code really well, even though your project may seemingly only require a subset of that knowledge. For example it’s not obvious adding an outlet could require rewiring the whole circuit and upgrading the panel to meet code.

Symmetrically, as you point out, one may find value in hiring a less expensive handyman to do their basic wiring, if you’re confident you just need a dimmer switch installed.

Well, as a contractor/consultant it’s such a waste of time when a company flips the Codility card. I simply walk away, it’s not worth the effort for a 6mo-1yr gig.