1) the product idea for something to build that you can sell profitably
2) the hiring skill
Neither of those are trivial to develop.
I do tend the think the second isn't that hard to develop for specific situations (FAANG interviews are really aimed at hiring people who they could swap around to large different sorts of teams or roles and get along with, and be productive with, a huge swath of the company, though IME in practice that doesn't happen terribly often these days). But if you're doing it for someone else's company, your upside is capped by the upside of the company you're working for. And your reward is too - you're not going to command 10x the pay from a CEO you report to who doesn't themselves understand the difficulty of the hiring you're doing compared to "just copy Google" (or why they wouldn't have good results with the latter).
The relative absence of numbers from the conversation is what allows many directors of football/soccer (those who recommend the purchase or sell of players) to boast about their stature as talent scouts.
Indeed, it is not surprising that directors of soccer focus on successes and tend to gloss over the many blunders made. Which makes it difficult, as in many other fields, to distinguish the role of luck and ability in successes and failures. Almost always, in my experience, those with the most money and at least an average "eye" succeed in doing the best that can be done: everything else is a game of chance.
A more sophisticated and technically oriented person would say that the most that can be done is to increase the odds of success, but then we are all left to the whims of fate: the young soccer player who many thought would be a sure bet will turn out to be one among many, and the average player at age 16 will have the growth spurt in the next two years that will make his body more conducive to stellar success (but was the growth spurt predictable?).
What I can say is that there have been several attempts to create a special agency for the identification and management of talent in the technology industry, but none has turned out to be more than a higher-level contracting firm.
It's luck of the draw. That's really how it's always been.
If you get bad interviewers, you're in for a bad time.
If you get asked questions you know, you're in for a good time.
It's possible your odds of getting bad interviewers has increased. But I doubt the increase is that substantial.
It's interesting to me how much stock these companies put into hiring, and then let countless terrible interviewers continue to interview.
You'd think they'd recognize high quality interviewers and get them doing more than 1 interview per week, and not let people interview until passing a pretty high bar in the quality of their interview.
Instead, they pretty much let anyone interview (as long as they've been at the company for N months and gone to some training).
I think you substantially over-estimate the eagerness of software developers to take on unrewarding admin work.
Why do you assume it can only be un-rewarding?
If hiring is so important, maybe it's worth paying the best people extra to get it done correctly.
If it's not that important, it'd be nice if they'd stop pretending like it is.
It’s an important day for them, so you owe them your full attention. Can’t be browsing the web or replying to emails.
After a decade of interviewing, there is no excitement in watching the thousandth interviewee make the thousandth attempt to count the set bits in an integer.
Hiring is important, it's just very very very boring.
Im not victim blaming by saying you may just be applying to the wrong places, am I
OTOH, when you are hiring for leaders (directors and above), past experiences and behavior questions are all you talking about.
Personally, I find enjoying the 8+ hours I spend at work each day more important than raw TC, so have made the tradeoff (and still have a reasonably high TC) which works for me.
That said, I would love a more laid back research programming job in academia, but that requires an additional 50% cut from where I am now.
It's still understandably frustrating that knowing how to write software seems largely irrelevant to getting a job at a FAANG.
Diving into a github and then asking questions about it, in the same way as you would ask about stuff on a CV seems like a massive win for the skilled interviewer.
Resume inflation is a big thing.
As for the CV that says took an online class that involved git cloning a repo and running scikit-learn on some toy examples is a CV you should discard as they probably just went for a cup of coffee and looked at the clouds for an hour.
I would add that I don't think you should think of it as "guilty-until-proven-innocent"; it's just the reality that you are competing with large numbers of amoral people who will lie to get a job. This is much less about culture match than you think.
Pretty much every other job in every other industry seems to do OK with education, experience and references.
If I want to hire a guy to tile my bathroom I don't need him to slap some tiles onto a whiteboard to get the job - some photos of the last few jobs he's done would be just fine.
Is the programming profession really so full of inept frauds? Is our industry so unable to regulate university education that qualifications are so worthless?
The main reason I hire the way I do is I learned that there are people who have resumes that look like mine (with 30+ years of coding) but they don't actually know how to do any coding, even just "please write out a simple class in python that does nothing". I go out of my way to give every candidate a fair shake and I'm not looking for some fast, set-in-stone rules. Nor am I looking for you to solve any really hard problems. Still it's been depressing just how many coders I interview who can't even explain how a hash table works.
IMO quizzing someone on whether they can solve an artificially tricky algorithm problem on the spot does not correlate very strongly the actual day-to-day work we do as software engineers. At best it's a measure of... whether someone can solve a tricky algorithm problem on the spot. It looks like engineering but it's not really what we do.
We're talking fizzbuzz-level verification, I'd say about 50% of applicants that I interview can't do fizzbuzz.
Do you want me to make it easier? how about reverse a string?
If you really perceive algorithmic live coding as torture, then it makes all the sense in the world to avoid it. "It's not for you."
But it's important to recognize two things: first that this is subjective. A different person may enjoy something you hate (I don't mind live coding for example) or at least won't consider is bad enough to be called "torture." So just because it's not "for you" doesn't mean it's bad.
The second thing is to recognize that there's a cost to being however we are. In this case, the cost of avoiding opportunities that require such interviews is that you may have fewer opportunities and they may be less lucrative on average. It's possible that (a) you are still able to find good-enough opportunities and (b) the downside is worth avoiding the torture. But you can also decide that this cost isn't worth it, and you have some room to shape what is and isn't for you, over time. For example, if the thing that makes this "torture" for you something like anxiety when being watched? If so, probably something like that shows up in other places in your life and maybe valuable to tackle anyway.
By no means am I suggesting that you need to do something different than what you're doing, just pointing out that these reactions are subjective, can have downsides, and are under our control to some extent.
> If you really perceive algorithmic live coding as torture, then it makes all the sense in the world to avoid it. "It's not for you."
I wasn't aware of this concept, but I can assure you that I don't have such a mindset. I tend to enjoy finding difficult challenges and complex topics to dig into which are way outside of my comfort zone. Leetcode practice is different, though, and yes, it does feel like torture to me and some other developers who tend to be rather quiet in general. While it's really tempting, I don't think I'm the right person to act as a community voice on this topic, although I do chime in sometimes.
When you're just starting this career path, it feels like homework. Heck, it's even exciting to solve problems which are a bit more complicated than whatever you're exposed to in school. However, try to spend several years in a row (8 in my case) trying and failing to pass such interviews (several in-person ones each year) and it will feel like torture. "If you'll spend a little bit more time on Leetcode, surely you'll pass next time..." No thanks! I'm done. The issue is not the lack of training, but the way in which these interviews are conducted by certain people which causes me distress and severe anxiety. I moaned about this before in various HN threads related to interviews and I might bother to blog about it some day.
> In this case, the cost of avoiding opportunities that require such interviews is that you may have fewer opportunities and they may be less lucrative on average.
Hasn't been my experience, but I'm not looking to maximise my income. I looked for advice on negotiation strategies and whatnot and felt happy with the offers I got. Others might have a different experience in this regard, but I think asking for less money in exchange for a psychologically-safe interview process is nonsense. At the end of the day, you get what you pay for and there are review sites anyway.
> For example, if the thing that makes this "torture" for you something like anxiety when being watched? If so, probably something like that shows up in other places in your life and maybe valuable to tackle anyway.
I think you're making certain assumptions here. I don't need to "tackle" my perception of what's safe and comfortable. We're all different and that's a good thing.
Most of the semi-successful companies don't need to be "the innovation machine".
The fitness function used by these companies is so broken it's not even funny. The ability to get things done and ship product doesn't even seem to be on their radar. They are optimizing for dead weight, which is how they end up with these periodic 1000+-person layoffs.
This it seems to be hard to standardize and streamline candidates validation based on their oss repos.
Also, afaik the goal of interview is to assess candidates thinking patterns, it’s hard to do quickly based on an arbitrary piece of code.
But it would also provide more valuable feedback.
1. Many (most) engineers are not in positions to be able to meaningfully contribute to OSS (e.g. all the code I write for work is proprietary).
2. It can easily be faked.
3. Personal projects can sometimes be interesting, but often don't showcase usefulness (unless front-end).
4. Many high quality engineers don't like to spend their free time doing more of their job.
Some repos, of course, are more impactful though. And I would probably say that if you are someone who loves to write code 16 hours a day, more power to you, you deserve to be hired! But many quality engineers would get missed if that is the desired bar (which is fine - other companies will just scoop them up).
(edit: OSS engineers extra upset by what I'm saying, but there is a lot more to this world than just what shows up in your GitHub feed)
Like I said, it would take more effort from the interviewers to dig into.
But it would also provide more valuable feedback and result in better hires.
> They're scaring off and filtering out excellent candidates with these excessive processes.
In my experience this is by design. The cost of a false positive (hiring someone they shouldn't) is far more expensive than a false negative (passing on a good candidate). Since FAANGs usually have a healthy pipeline of candidates and many who fail will eventually reapply they're okay with passing on possibly qualified candidates.
> None of them even cared about my Github repository that's full of code I write for fun.
As an EM at a mid to large tech company who has hired plenty of engineers I think ignoring people's Github is the right call and I don't even bother looking at candidate's Github unless it's for something like an intern role. Ignoring the fact that I don't know where the code on Github came from (did someone else write it? Is this just following a tutorial?), the code you write and the skills you use when you're writing code at home alone on a passion project is VERY different than the skills you need to be successful in a large, slow moving, bureaucratic corporation.
The cost of a false positive is 0 - they always hire the non optimum candidate there is a massive margin of error.
We all like to pretend that without our unique special skills the entire company would dissapear overnight, but it is a fantasy.
Context: I've interviewed around 200 people for our current org, and have "settled" on an interview that's used for a certain position. I don't think I'm necessarily good at interviews, and I'm not sure I'm "doing it right". I give the number to show that my sample size is reasonable. We're solving practical problems, with code. Our interview sessions are 45 minutes for each person (which is the fundamental limitation).
I don't look at GitHub projects, or really ask about previous work.
I used to, but I very quickly found that the vast majority of people lie on their resumes. Generally, the more senior, the more lies. Most people put bullet points of "achieved x" which end up really being "kinda helped those people who were achieving x". For GitHub, in most cases, it was obvious the person was not the author of the Github repo they gave me. They could answer questions about the code, but couldn't go into any depth about it.
I found that going over previous work, or Github repos, ended up taking a huge chunk of the 45 minutes, that put each person on a different standard, somewhat of their own crafting, leaving no time to find out if they were actually able to write a for loop or not, which I've had Phd candidates, with "15 years of experience", fail to do. For feedback of the candidate, I need a tangible justification, and clear demonstration of why I say yes/maybe/no. There are no "gut feelings" allowed from the panel; you need reasons.
So, I give the candidate a simple, practical, work relevant, problem to solve during the interview. No compilation, no running code, no worries about perfect syntax, and no esoteric knowledge/libraries required. They're free to use whatever language they want. It's an interview where they have to demonstrate their thought process, how they question/navigate the requirements, and ability to see consequences (which I think is 90% of programming).
Maybe a few percent have finished within 10 minutes. For these people, we had fun talking about their other work, and they showed great knowledge, as expected. But, as desired, that wasn't required to do first. The interview itself proved that they would already.
I would love feedback about this approach, even though it seems to be working well for our orgs needs. Ideally, I would take the Jim Keller approach, and have a nerd out session, but, I have a feeling people are more properly filtered before they would get to him.
With regards to Github projects and the like, I do ask about them in interviews, because I think there's a place for them as an early screening. Every engineer worth their salt has some war stories, and that's a good angle to approach their GitHub profile: if they have an interesting sounding project, asking them to talk about any interesting problems they ran into during its creation tends to make it immediately clear whether they're full of shit or not. It's not a great positive signal, but if they're not chomping at the bit to talk about engineering challenges, it's a strong negative signal.
Yeah, this is the biggest problem I have with it. I try to be conscious of this, and try to keep things pretty lighthearted. If I notice someone getting nervous, a "you're doing great for time" usually calms them. If not, I do follow up interview, if they did well seeking an understanding the problem, and could explain it verbally. I've had a few people pass with flying colors after a second round.
> and that's a good angle to approach their GitHub profile
I disagree here, with the (probably wrong) assumption that not having a GitHub profile adds a negative bias. Some people put their coding hours into work, with different lives outside of coding. All of the colleagues I've known with impressive GitHub profiles burnt out/left. Battle stories can be within corporate repos.
If the candidate is lying, just move on to the next.
Terms like "good thought process", etc aren't allowed in our feedback, unless specific examples are given. Terms like "team fit" will get your feedback outright rejected. No mushiness allowed.
I recently had an interview where, for the first time, the coding round was Coderpad based, which allowed me to run my code and build my solution bit by bit with instantaneous feedback.
I know for sure that I wouldn't have been able to solve that recursion problem without that live terminal. I now am going to insist on this live terminal in all of my future interviews and refuse to do any that don't employ one.
I think it helps my anxiety immensely.
Regardless, I'm going to offer that as an option and see how things go. Maybe everyone will be in a better mindset, allowing me to add a bit more complexity/depth to the problem.
That's like saying that janitor jobs ripe are for revolution/innovation. You would think I'm joking, but most companies really do not care, despite what they say.
And I've failed FAANG-ish interviews where no amount of leetcode grinding would have helped me.
In my opinion, it's getting rather silly. They're scaring off and filtering out excellent candidates with these excessive processes. None of them even cared about my Github repository that's full of code I write for fun.
Software developer hiring is definitely ripe for revolution/innovation.