Interview Bar is Going Higher
Just interviewed with several big/unicorn company. Feel like the interview bar is much much higher compared to several years ago. I practiced several hundreds algo questions several years ago, and can easily nailed offer from all major companies. Right now no matter how many leet code questions you have practiced, you are still highly likely meet with hard level questions you never saw before.
Demand is great, tons of folks now know that high TC is a thing so the competition is fierce.
Lol. I think it’s people like us who practice 100s of problems before interviews with the expectation of seeing them in the interviews are making interviews hard for everyone.
Exactly. Recent hires tend to give more difficult problems making interviews harder and harder.
Luckily even tier 3 companies pay very well by any objective standard
Aside the stock
Lol... What is your objective standard? FAANG+M is one tier above all others when it comes to TC.
Thanks to Leetcode based interviews
Can confirm Winner's Curse.
Don’ think it is going higher. The level of candidates is getting lower.
How so? The number of CS grads from stanford is like 10x greater than a few years ago
That’s exactly what I mean. And 100x greater than a few years before that. They cannot all be as good as Larry&Sergey. Lots of folks going into CS because of high pay, not because this us what they love doing.
The problem is recently hired folks tend to set the bar higher thinking he himself could pass the same bar but they are giving harder problems with stricter criteria than when he passed. This keep increasing the bar. Another factor is wide use of leetcode. This is evident for companies with smaller set of problems such as airbnb. Their expectation is much higher because many of candidates have already practiced the same problems thus solve them quick with clean code. You are just competing against other candidates. Another thing is engineers have strict criteria than possible any kind of professions. It is our nature to find a small thing and be picky about them. At onsite if you let only 1/4 of candidates pass, that can be disastrous. You think this candidate will be able to pass others since he demonstrated his skills. That is rarely the case since it depends on whether the candidate solved the problem before. His chance may be slightly better but even if it is 1/2, chance to pass two interviewers is as low as 1/8. It is simple math. The candidate would have better chance but it won’t be 100%. Even if you multiply 3/4 to that, that is 3/32 which now is close to 10%. This is assuming your company gives an offer even though he fails the next one. Normally when people get into highly competitive companies such as Google or FB, they have a chip on their should to filer out bad candidates. That is how their onsite passing rate comes down to 2-5%. You think these people are smart. They are smart at very narrow spectrum and can’t see the big picture nor have will to improve the practice.
You can not set the bar arbitrarily high even if you are FAANG. Google alone hired 20,000 last year. You cannot raise the bar if you are hiring more people. Individual interviewers can’ change this by asking all LC hard problems. If you are not FAANG then you have to hire anyone with a pulse. At Oracle I was pressured to give hire ratings to unqualified people “because we are going to lose the req”.
I wonder if the quality of hires is going up as the bar goes up. Winners curse suggests that the worse your screening process the worse off you will be when you win a candidate. These companies started using LC to screen out false positives, but that means it doesn’t make sense to make LC style interviews harder since you were already filtering false positives and making it harder filters out good candidates. https://en.wikipedia.org/wiki/Winner%27s_curse