Story time.
A few years back I was in hiring loops as an interviewer. Half a dozen engineers in my company interviewed a candidate during a morning. Everyone had their notes and we were discussing if we should hire the guy or not. Everyone gave thumbs up except one engineer. It turned out he tested the candidate on data structures and algorithms, and the candidate hadn't memorized the computational complexity of some bullshit algorithm that only shows up in lame trivia games no one cares about. Based in that, that particular interviewer gave thumbs down. We discussed everything, went through all notes, and ultimately based on all feedback the hiring leader gave a thumbs up. The data structures engineer threw a tantrum. In his opinion, we would be lowering the hiring bar if we hired a candidate that failed to meet his bar on what he felt were basic competences. The hiring leader asked how many times he used that algorithm in his everyday work. The data structures engineer answered that alas he never used it at all, but still he felt that everyone working in his org should know that.
I feel some companies don't have good hiring managers and give free reign to bad engineers who have no idea what it takes to deliver production code.
I have worked on teams trying to improve our interviewing process, and one key goal was aligning everyone in exactly what technical and personal attributes we were trying to hire for.
Even so, the data structure engineer cracked open Trivial Pursuit on the candidates and wanted to reject him based on one answer alone.
Of course if you're looking for a software jeopardy contestant maybe memory recall is important, but for the rest of us it is much better to have a good understanding of fundamentals and know where to look when it comes to specifics!
In my anecdote, the data structure engineer was indeed tasked with evaluating the candidate on his data structures knowledge. Except that the candidate had guidelines that went in the line of assessing basic understanding and proficiency in following an algorithm, but instead opted to make it a game of Trivial Pursuit.
It's always trumped up nonsense blown out of proportion. In one debrief, they made a stink because they botched an implementation of a red-black tree. it was the most pretty, trivial nonsense. The only time people are implementing red-black trees by hand are... when they're studying for our bullshit software interviews.
We also had a guy with a Ph.D in.. something. He wanted everyone to know he was very smart. Especially the interviewees. The problem he asked was so bizarre and complex that it took those of us already at the company several hours to figure it out upon giving it a try. it's just, like, dude, what is the point of what you're trying to do here? How small is your ego that you need to bully strangers on irrelevant nonsense?
For what it is worth; based on what I've seen in interviews. Programmers can only rank people who are close to their own level. If someone like Terry Davis of TempleOS fame shows up for an interview, should the interview declare them as a madman? a genius? both? If we need to build an OS, are his design choice eccentricities sly moves of genius or evidence of a diseased mind?
It isn't as much that strong programmers are a threat (that might be a factor) but more that big gaps in though lead to both productivity and inscrutability. And people avoid things they don't understand.
You can google many things easily. Understanding the answer and knowing how to apply it to different problems matters a lot more. I still expect some concepts and definitions to be remembered if you’re claiming expertise in a topic.
that's wrong?
I hate those types of trick questions.
Yes, they can Google the answer in 3s.
Yes, I could also ask a long-form question about how they would approach wrapping an IO interface in Python.
But if they don't understand `with` or resource/context management, the chances that they would produce a reasonable answer to the long-form question are tiny. They will produce an answer anyways, and it will likely be a waste of time. So we cut to the chase and target the core concepts directly.
Heck. Sometimes going into topics “not relevant” for the job is revealing. If their cv says they spent 5yrs writing Linux dev drivers but the position is to write golang web services I may still ask about device drivers if I can get a domain expert. Good people tend to understand very well the things they’ve worked on.
I would suggest anything on the applicant’s cv is fair game. What makes no sense is asking about removing duplicates from a linked list when that’s nowhere near what the job will entail.
Now trying to remember a specific algorithm like red-black tree rotation or Djikstra’s is something completely different
I specifically try and answer their questions and “hold back” a bit when answering if I can. I also sometimes will drop a nugget of knowing more and study their reaction. Some interviewers react quite poorly if you hint you know considerably more about a subject than they might want you to know.
I have friends with a lot more education than myself (PhD, masters, very senior) and they say they have to do this A LOT.
The goal is to see how the interviewer handles not being the smartest in the room about everything. This is a big deal because it can suss out whether the other person is “on good behaviour for the interview” but otherwise has an enormous ego during regular work hours.
No one wants to work for an ego maniac vetting out all the A+ candidates out of fear of competition.
My advice would be to not play mind games and just interview as yourself. Interviews go both ways, so if your interviewer is a psychopath you'd want to know sooner. Then just give the job a pass.
However, my experience with FAANg companies is: the interviewers cannot handle an interviewee who knows more, has more experience, and who comes up with better solutions to their stock 'whiteboard' problems.
Such a person should not squander his/her talents in a FAANG. That's why there are start-ups.
I will happily hire someone who’s a threat to take my job. If I don’t, I can’t ever move up either.
Yes, I know the answer is to keep them at market wages and don’t fall into the salary compression trap.
But often that isn’t under your control as a hiring manager. HR won’t give you a budget for raises. But will give you the budget to hire at market rates.
If they know the technology that’s one less thing to worry about.
“what the hell were they thinking using X like Y!”
“Why didn’t they use Y here”
“Oh man, this looks like monkey patched code from 2 full versions ago, the right way to handle this now would be 1/3 the code, but oh they can’t update it because they’re mutating the data structure in order to Z… Fuck…”
With maturity, you realize whatever they did worked well enough to get the company to a position where they were able to hire you.
As Corey Quinn (a well known “cloud economist”) says, “legacy software is software that generates revenue”.
My point was exactly the opposite and parallel to what you said. But even with experience it takes time to decode those decisions.
The point is that it takes time to understand the whys of code that is counterintuitive at first.
Even if you are a 100% match for the technology.
When I joined a kernel gfx driver team, the expected ramp time was ~6 months. My friend working on Mesa had a >1 year ramp time!
If I open the source for a Django or React app, I can get to fixing bugs in a few days or less. It just depends.
“If you need someone who can hit the ground running…” then you need to have documented the crazy decisions that led to the awkward implementation that is your product. I have yet to meet a company who is using standard, recommended, tried-n-tested, well-known solutions to typical problems. Unless you write docs or stick to standards, no outsider has the context to “hit the ground running.”
I learned to do this when I was working on my master's, and my advisor suggested a project to me, and when I asked him for help figuring out where to start, and he told me to read through the code until I understood it, and come to him if I spent more than an hour or two stuck on things. Thanks, Mikhail - your advice was fantastic.
The best advice I've heard is "start from the data structures, they are at the heart of everything".
I like inventing purposes while reading through code if I don't have a real one. I call it "scavenger hunting" and often use it as a programming exercise when teaching. For example here is a scavenger hunt for CPython and Postgre - which lines of code would you need to touch to achieve these purposes?
https://github.com/python/cpython In json module: add parameter to load() and loads() that when True only allows spaces as whitespace characters in the json (and not tabs or newlines)
In re module, python regex has syntax for a digit (\d) and for an alphanumeric or underscore character (\w). Add \l that means "any letter".
In re module, python regex has syntax for hex escapes (e.g. \x0a), octal escapes (e.g. 0o012), unicode escapes (e.g. \u000a). Add binary escapes (e.g. \q00001010)
In python C implementation: Replace the hash function used when inserting a string key to a dictionary
In python C implementation: Add an average() builtin function (like e.g. max(), sum())
https://github.com/postgres/postgres Better string hashing algorithm for query plans that involve hashing a column
"Better column combination hashing algorithm for query plans that involve hashing a column
Some queries will want a hash on (a, b) which involves calculating hash(a), hash(b), and then combining them into combination_hash(hash(a), hash(b))"
"Add a new kind of step to execution plans
e.g. a step that logs to a cloud logging service. No need to compile it into queries, just to add the data structure for it and code that runs it when relevant so it could be compiled"
"Add gzip compression as a way to avoid page splits
Today when a page is about to be split, bottom up deletion and deduplication are attempted to keep the page from splitting. We want to add GZIP compression as a third way to try and avoid expensive page splits."
"Add a complex number column type
The column should hold a float for the real part and a float for the imaginary part"
Add Configuration Flag to Never Use Covered Index Optimization
"Add new type of index
e.g. index that runs Principal Component Analysis to index high dimensional data, but don't care about the details, just the integration points"
I've delivered every single time, but I can't do this shit more than a few times a year; I don't say this lightly: it's fucking unhealthy..
Tech is usually easy to pick up, and highly company specific. Any good engineer should be able to learn new tech, and things will change over time.
Culture is much harder to "teach". For me testing for culture fit has become much more important than tech acumen, particularly at more senior levels. You need to have a sense of at least whether the candidate aligns with the company culture, and is willing to fit.
Not doing this risks several scenarios:
a) The candidate takes a while, but finally fits with the culture. This is suboptimal but ultimately a good outcome.
b) Candidate refuses to fit with the culture. This will likely cause performance issues, and the employee will ultimately leave or get fired. Not a good outcome.
c) Candidate actively subverts the company culture. This is the most destructive outcome and can be quite damaging, particularly at higher levels.
Checking that someone will be a good fit for the company is better for both the company and the candidate.
Counter note: I've done 100s of interviews. A bro, bruh, or brah wouldn't even register as something notable to me. This is just how my younger coworkers talk.
Their use of "bet" has taken some getting used to.
Without wanting to get into a discussion of the gender-neutrality of those words, I'm not a man and everyone that I talked with there was, and it just felt odd.
Note to everyone here: Not everyone is as uptight as this guy and you shouldn't be either.
Sometimes two philosophies aren't necessarily wrong but simply aren't compatible. Choices like emphasising frequent collaboration and constant communication or async-first deep work. Mismatches there are genuine culture misfits and it benefits everyone to realise that early and move on with no hard feelings.
But a lot of ideas in how a team works are not so black and white. It's valuable to hire people who have different ideas and aren't afraid to advocate for them when the time is right as long as they are still team players who will accept a result that isn't their personal preference if that's the team's consensus or management's direction. Not hiring people like that because their personal preferences don't align with the team's current way of doing things is also a culture misfit but in this case it's a dodged bullet for the good candidate.
> On the other hand if you never hire anyone with different ideas then you'll inevitably create a monoculture that had difficulty seeing its own weaknesses and making improvements.
This is a balancing act for which there isn't a hard and fast rule, because organizations need a culture, but not all cultural values are good.
A group might want to bend on someone who is weaker on their Documentation And Testing value, but not bend on someone who is weaker on their Honesty value; but maybe they won't bend on Documentation And Testing because they view it as essential and don't want to risk that the FNG won't internalize it.
"But that isn't the cultures I object to," I've heard critics say. Of course not. But sometimes those are the cultural values that folks are trying to conserve. That's why there's no hard and fast rule.
Sometimes the cultural values are dishonesty or bigotry or cronyism, or simply an aversion to being called "bro". But often they're honesty and open-mindedness and fairness. We should not toss out the baby with the bathwater.
Maybe only about 10% of people who are sadists or calculating assholes actually follow that motivation? Unless you had a very special sample pool. Most interviewers i've met - both good and bad - were just trying to do what they think is right, with varying degrees of competence.
The best interviews I’ve ever taken and in turn, host, are the ones where we break down a problem with low code, no gotchas, little examples of specifics just to make sure they know how to write it in language of their choice. That they understand the problem, can satisfactorily break it down into actionable tasks, and come up with a solution to the problem.
Everything else: tech stack, jira, agile, culture - is company specific and should be taught to your employees and not be made the bar. If you need someone who can hit the ground running you can focus on a tech stack but ultimately it’s up to you, the company, to ensure they fit. Not the new hire.