|
I'm a bit lost myself to be honest. Are you expecting them to fix the tokens and then feed it into a standard JSON parser or to write a complete JSON tokenizer and parser right there in the interview?
|
#
?
Oct 9, 2023 16:43
|
|
|
|
| # ? Apr 27, 2024 11:31 |
|
|
Sapozhnik posted:I'm a bit lost myself to be honest. it appears to be "parse this string-represented data structure, some how some way, just do something and print something out." it's 2023, this poo poo is built in. even golang does stuff with the range keyword. this is a completely trivial ask, if somebody can't get a dozen lines down and talk about builtins vs using stacks, or whatever they're probably pretty bad. i kinda fail to see how you can get "im so confused do you want me to write a complete json parser" out of such a completely boring, bog-standard interview question. plus there's that whole "interviewees are supposed to ask interviewers clarifying questions" thing which, well, surely you weren't expecting a complete description of the interview question on the shitpost forum, were you? op is venting, not asking you to solve the thing
|
|
#
?
Oct 9, 2023 17:41
|
|
|
It’d be more like - We’re Subway, we have a sandwich order taking system - someone wrote their own sandwich notation before JSON ever existed code:Size: Large Toasted: true Toppings: Turkey Toppings: Cheddar And things like code:And I’m hoping they get this far code:E: Anything lighter weight would be something they already know like a CSV parser. and I’ve had to convert a home grown pipe-delimited format into something usable. Coffee Jones fucked around with this message at 18:09 on Oct 9, 2023 |
|
|
#
?
Oct 9, 2023 17:55
|
|
|
for me it’s not confusion, it’s that it’s the kind of interview question that superficially seems like an actual job task (ingest and transform some input to presumably solve a business problem) but is actually something completely different (just write some bullshit to perform a very specific particular transformation on a very tiny input data set that the interviewer may or may not be generating adversarially, and then talk about the bullshit.) “talk about the bullshit” is the actually important part of the process but you have to first produce it so you can talk about it, and reminding myself that the interviewer doesn’t actually care about most of the things i’d be thinking about if i were doing this in an actual job is difficult for me.
|
|
#
?
Oct 9, 2023 17:59
|
|
|
Ah okay so it's asking for something that accepts a particular ad-hoc string representation of a particular data-structure, not something that can accept any JSON-like document. So yeah assuming that escaping isn't a problem you'd split it along the various magic characters and then do a switch/case on the field labels or whatever, then apply some hardcoded set of validation rules after you reach the end of the input.Achmed Jones posted:it appears to be "parse this string-represented data structure, some how some way, just do something and print something out." it's 2023, this poo poo is built in. even golang does stuff with the range keyword. this is a completely trivial ask, if somebody can't get a dozen lines down and talk about builtins vs using stacks, or whatever they're probably pretty bad. The way it was phrased, to me, implied "here is a single example line of a completely general language whose grammar is broadly equivalent to JSON, write something that can parse it", specifically using the verb "parse", and the keys and values are very generic and not specific to any real or imagined application so again it reads like an extremely under-specified description of a generic serialization format, unlike the Subway clarification given later.
|
|
#
?
Oct 9, 2023 18:19
|
|
|
so yall might remember me from such hits as "use a rubric" and "use a rubric 2: you loving fucks, just write down success criteria" anyway i was talking with a friend the other night and he's the type to put hours into reading academic literature on this stuff, and i am gratified to know that according to the literature on structured interviewing: 1) "just having a conversation" does worse than a coin-flip at hiring 2) rubrics are super important, and need to include anchored rating scales 3) other basic well-duh-if-you-think-about-it-for-30-seconds things for reducing bias like "dont talk to other interviewers before they interview," "make sure the interviewer understands the problem," "take notes," and "provide training to interviewers" are very important too
|
|
#
?
Oct 9, 2023 18:19
|
|
|
Not a Children posted:The only two people I've ever seen terminated for underperformance were doing close to zero work for a span of > 12 months Perception of zero work mattere more than zero work here. I've seen people maintain perception of work with zero or negative output for years.
|
|
#
?
Oct 9, 2023 18:20
|
|
|
Sapozhnik posted:The way it was phrased, to me, implied "here is a single example line of a completely general language whose grammar is broadly equivalent to JSON, write something that can parse it", specifically using the verb "parse", and the keys and values are very generic and not specific to any real or imagined application so again it reads like an extremely under-specified description of a generic serialization format, unlike the Subway clarification given later. i mean yeah, but...yospos, not wanting to spray literal questions all over the internet (or necessarily provide a whole new example) etc. you gotta read it with some charity. and since homeboy was complaining that interviewers didn't know how to map/split/etc, it seemed pretty clear that we weren't in "now implement a red-black tree on paper, but it has to compile" territory. anyway interviewing sucks on both sides
|
|
#
?
Oct 9, 2023 18:21
|
|
|
leper khan posted:Perception of zero work mattere more than zero work here. I've seen people maintain perception of work with zero or negative output for years. how? asking for a friend
|
|
#
?
Oct 9, 2023 18:22
|
|
|
Achmed Jones posted:how? asking for a friend Only ever product tangible things that are completely hosed. Only do this infrequently. On at least a weekly basis, produce convincing sounding non-tangible items. Develop rapport with your manager and peers; be able to answer and resolve questions in a way that sounds like you're helping them but in practice is not actionable. I can't provide help on how to actually implement that unfortunately. Which is potentially the first step. 🤔
|
|
#
?
Oct 9, 2023 18:28
|
|
|
Achmed Jones posted:so yall might remember me from such hits as "use a rubric" and "use a rubric 2: you loving fucks, just write down success criteria" does the literature suggest a way to update the rubric when flaws are found without undermining the consistency that allows it to function as a rubric? i’m fully on board with Use A Rubric but the resolution to that tension seems non-obvious
|
|
#
?
Oct 9, 2023 18:45
|
|
|
Achmed Jones posted:"provide training to interviewers" What does good training look like?
|
|
#
?
Oct 9, 2023 18:50
|
|
|
one of my criteria for interviewing are: - what do you spend your time on in a week? (might include poo poo like: reviewing code, patching issues in some legacy code base, figuring out messy requirements, choosing how to prioritize poo poo as schedule slips, meetings, updating tickets, clarifying what people mean) - what is the work you do pretty much every month or so? (writing docs, updating deps, sending some reports, some white board architecture, etc) - what do you rarely do but wish you did more of? (probably algorithms, optimization, etc) chances are 80% of this industry’s interview time is structured on that last category because that’s what engineers wish their job was and that’s where pride and professional worth comes from, whereas the vast majority of the actual job is spent elsewhere and we keep collectively filtering for the most irrelevant poo poo, or poo poo that’s relevant like twice a year on average. of course this can vary by role, seniority, and industry your product targets. but either way a good interview process should ideally spend more time on basic day-to-day poo poo because that’s what you’ll do most.
|
|
#
?
Oct 10, 2023 00:18
|
|
|
one my current job's questions is "here's some code, please review it" which i really liked
|
|
#
?
Oct 10, 2023 00:51
|
|
|
i would like a new job where i am paid large amounts of money to nap and snack all day while doing no work. please advise how to interview for this role
|
|
#
?
Oct 10, 2023 00:53
|
|
|
re: training, idk. start with your best effort and revise. i suspect that just telling people what you're doing and why and letting them practice does a whole lot of good. but feel free to search jstor. if an hour of conversation and a couple hours of reading taught me anything, it's that giving a poo poo and trying is like 80% of it re: updating the rubric, idk about the literature but having made and used a lot of rubrics, you just update it. it's pretty simple and it doesn't take long before it gets good enough updating is pretty rare. if you're confused a lot and find yourself getting a lot of answers that don't match any of your examples, it could be that your question sucks, that you need a more fine-grained rubric, that it's too fine-grained, etc. i suspect there exists a body of knowledge here, but it doesn't take long to get a feel for it in my experience. pm me if you want to talk about problem cases and stuff, can't really get into too many specifics on public forums but pms might work better; i'll help if i can
|
|
#
?
Oct 10, 2023 01:17
|
|
|
nudgenudgetilt posted:i would like a new job where i am paid large amounts of money to nap and snack all day while doing no work. please advise how to interview for this role position yourself as an “individual contributor”
|
|
#
?
Oct 10, 2023 03:30
|
|
|
raminasi posted:one my current job's questions is "here's some code, please review it" which i really liked “Please don’t just send me code reviews without first giving me proper context. I don’t own this tool.”
|
|
#
?
Oct 10, 2023 03:34
|
|
|
raminasi posted:one my current job's questions is "here's some code, please review it" which i really liked this is a good strat imo
|
|
#
?
Oct 10, 2023 07:03
|
|
|
Kazinsal posted:position yourself as an “individual contributor” this isn't enough on its own, you gotta be at a high enough level that your job is some combination of unquantifiable technical advice to others and inscrutable personal projects that nobody else expects to understand. this is called "staff engineering." tk posted:“Please don’t just send me code reviews without first giving me proper context. I don’t own this tool.” lol iirc my first question was "before i read this tell me what it is and why i'm reviewing it" and they actually had answers for me, which i interpreted positively
|
|
#
?
Oct 10, 2023 07:12
|
|
|
nudgenudgetilt posted:i would like a new job where i am paid large amounts of money to nap and snack all day while doing no work. please advise how to interview for this role this is my (and my boss’s) job find a megacorp that makes too much money on a dumb but in demand product and a team just isolated enough to survive throughout the years of layoffs and reorgs
|
|
#
?
Oct 10, 2023 11:41
|
|
|
remote contracting for enterprises who have no clue what they’re doing also works
|
|
#
?
Oct 10, 2023 12:00
|
|
|
one day I should work out if I have anything of value to offer someone beyond a legal right to count drugs the problem - is that I am a clown
|
|
#
?
Oct 11, 2023 01:52
|
|
|
echinopsis posted:one day I should work out if I have anything of value to offer someone beyond a legal right to count drugs More of value than any of the rest of us can offer.
|
|
#
?
Oct 11, 2023 03:04
|
|
|
i have to count drugs illegally
|
|
#
?
Oct 11, 2023 04:28
|
|
Achmed Jones posted:. anyway interviewing sucks on both sides Yup. Too hard and no one passes. Too easy and “What’s the difference between Integer and int?” “I’ve been doing react for a year but don’t know what package.json is for.” “Every contribution I make just creates more work for you.” Too generic and there’s people who grind leetcode. If you’re Amazon you put people through an automated tester - here’s the problem, pass these unit tests, we won’t tell you what the tests are. It’s extremely impersonal and mostly just pass / fail . If you use an interviewer service like Karat, less impersonal but interviewer is just a 3rd party person with nothing to do with the job … also expensive. Just had a guy pass the funky json question. I’m feeling better. Asymmetric POSTer posted:this is my (and my boss’s) job Something super niche but necessary like veterinarian office or mortician software
|
|
|
#
?
Oct 11, 2023 07:38
|
|
|
leper khan posted:More of value than any of the rest of us can offer. but not in a remuneration kinda way
|
|
#
?
Oct 11, 2023 08:03
|
|
|
echinopsis posted:but not in a remuneration kinda way if you can count pills legally for a little money, you can count them illegally for a lot of money the more you know!
|
|
#
?
Oct 11, 2023 08:07
|
|
|
we just had that one peep who got a nz job offer and it was absolute dog poo poo so you might have to move to a figgieland unfortunately for that remuneration to be big
|
|
#
?
Oct 11, 2023 08:12
|
|
|
i’m on 75k of clown dollars but tbf I only work 4 days a week  my partner earns more than twice as much as me lol. still life ain’t cheap bob dobbs is dead posted:we just had that one peep who got a nz job offer and it was absolute dog poo poo so you might have to move to a figgieland unfortunately for that remuneration to be big lol what do you consider dogshit i’m too much of a jack of all trades to be useful figgie wise. people like working with me, and as their superior (half the time i’m at work) people appreciate how I am respectful and kind and always willing to help and answer questions.. I never get annoyed or pissed of and do my absolute best to make sure no one feels bad. but none of those are really that useful to me for getting more big bucks soz I am not trying to be down on myself like those are great qualities and people super appreciate me.. but I need to figure out the next move. now that i’ve got my adhd mostly sorted maybe it’s possible i can get ahead 🤔 echinopsis fucked around with this message at 08:35 on Oct 11, 2023 |
|
#
?
Oct 11, 2023 08:23
|
|
|
like 75k nzd lol levels.fyi for medians in figgieland. i actually think nz and sf real estate prices are quite comparable, dunno about other costs of life or australia if you dont wanna move that far. remote jobs had a lacuna there for like 2 years where they were for juniors too but at this point theyre mostly for seniors again. its also not completely clear that theyre such a good deal for juniors junior hiring is not in a great state rn bob dobbs is dead fucked around with this message at 08:30 on Oct 11, 2023 |
|
#
?
Oct 11, 2023 08:25
|
|
|
no chance of moving my only chance is somehow leverage my front line experience. how I don’t know but I imagine it’s possible
|
|
#
?
Oct 11, 2023 08:36
|
|
|
have you considered becoming a roadman?
|
|
#
?
Oct 11, 2023 08:38
|
|
|
as in just wandering down the road?? yes
|
|
#
?
Oct 11, 2023 08:57
|
|
|
i would fail the funky json test because i struggle to solve problems where the question is "we arbitrarily hosed a thing up and you have to fix it" and the answer is just tedious and not interesting at all
|
|
#
?
Oct 11, 2023 15:47
|
|
|
Corla Plankun posted:i would fail the funky json test because i struggle to solve problems where the question is "we arbitrarily hosed a thing up and you have to fix it" and the answer is just tedious and not interesting at all i have bad news about working in this field
|
|
#
?
Oct 11, 2023 15:49
|
|
|
Corla Plankun posted:we arbitrarily hosed a thing up and you have to fix it i hope you never have to work with javascript and famous inventors of wheel 2.0
|
|
#
?
Oct 11, 2023 16:34
|
|
|
Corla Plankun posted:i would fail the funky json test because i struggle to solve problems where the question is "we arbitrarily hosed a thing up and you have to fix it" and the answer is just tedious and not interesting at all i mean it's either "grinding code that's tedious and not interesting" or it's "completely unrelated to the actual job" but it's not both, that much I know for real though, are you saying that you can't pass an interview where you don't, personally, find the question to be interesting and engaging? and i assume you can't generally do work on things that aren't interesting and engaging? that sounds like it would make things very difficult, and also tremendously blur the lines between 'work' and 'hobby,' which i dont think i could handle. Achmed Jones fucked around with this message at 16:39 on Oct 11, 2023 |
|
#
?
Oct 11, 2023 16:37
|
|
|
my favorite interview question was "implement an api for excel" and it went like this: Q: implement an object that acts like an excel spreadsheet and supports the following: set(place, value) and then as the interview went on it got more and more complex in increments: get(place) set and get where the set value can reference other places (e.g. '=A2+B2') a longer interview could get into row-wise or column-wise stuff if need be It was fun and an easy way to demonstrate that I knew how to write python and it was easy to discuss/reason about Achmed Jones posted:i mean it's either "grinding code that's tedious and not interesting" or it's "completely unrelated to the actual job" but it's not both, that much I know i think the "point" of the interview is important to call out: I'm happy to do a million interviews where the idea is just "do you know how to write python and reason about code" but if the point of the interview is to weed out people who are bored by tedium this interview is correct to weed me out. I've got a decade of xp and a masters degree and the Correct Answer to an on-the-job problem of "we have an interface that is hosed up javascript" is always to negotiate a better interface with the people who are providing it and never, ever to try to guess the syntax and homebrew a parser for someone else's bad design choice
|
|
#
?
Oct 11, 2023 17:02
|
|
|
|
| # ? Apr 27, 2024 11:31 |
|
|
Armitag3 posted:i hope you never have to work with javascript and famous inventors of wheel 2.0 oh yeah i dont know a single framework on purpose and don't have any references to javascript/html/css on my resume unless i'm trying for a job that lists them as a nice-to-have and I really, really, really want it
|
|
#
?
Oct 11, 2023 17:04
|
|