I got kicked out of Columbia for taking a stand against LeetCode interviews(linkedin.com)
linkedin.com
I got kicked out of Columbia for taking a stand against LeetCode interviews
https://www.linkedin.com/posts/roy-lee-goat_i-just-got-kicked-out-of-columbia-for-taking-activity-7310834407433453568-tqAm
18 comments
How terrible it would be for someone (or some company) to waste the time of 6 people (or 600 people) in an unnecessary and ultimately useless interview process.
Turn it around. Do these companies pay every developer they interview, including the ones they reject for the bullshit they put candidates through?
Sorry mate, but as a hiring manager, spending time to talk to people for fit assessment is part of the cost of doing business. You don't get to call it theft because the guy wasn't game to play your games. You're the one erecting barriers, not them.
Sorry mate, but as a hiring manager, spending time to talk to people for fit assessment is part of the cost of doing business. You don't get to call it theft because the guy wasn't game to play your games. You're the one erecting barriers, not them.
Why the hate for leetcode testing?
From what I hear from friends at Leetcode employers, they have barely any developers who can't develop for shit. That is better than average.
'Barely any developers who can't develop for shit.' That's a frustrating sentence to unpack.
Is the claim that the majority of developers are proficient because of Leetcode as a filter?
I don't think Leetcode is a useful interview tool when the actual role deviates from the style of questions that are expected to be completed in an interview scenario. And this is a prevalent issue.
Clearly it can be gamified and I don't think this makes for good engineers. It makes for people that are good at solving Leetcodes. Jira tickets aren't Leetcodes.
Whatever technical filter is implemented (pairing/take home/in house white boarding etc.) I care much more about observability, logging, approach to difficult problems etc. and general attitude.
Run me through the architecture for previous solutions that have been worked on. Did you own it? Were there any outages? What happened when requirements change? How can you ensure what you've created is fit for purpose for the organisation?
There's so much more to engineering rather than just code. And in my experience many engineers are worse at this than they think.
Is the claim that the majority of developers are proficient because of Leetcode as a filter?
I don't think Leetcode is a useful interview tool when the actual role deviates from the style of questions that are expected to be completed in an interview scenario. And this is a prevalent issue.
Clearly it can be gamified and I don't think this makes for good engineers. It makes for people that are good at solving Leetcodes. Jira tickets aren't Leetcodes.
Whatever technical filter is implemented (pairing/take home/in house white boarding etc.) I care much more about observability, logging, approach to difficult problems etc. and general attitude.
Run me through the architecture for previous solutions that have been worked on. Did you own it? Were there any outages? What happened when requirements change? How can you ensure what you've created is fit for purpose for the organisation?
There's so much more to engineering rather than just code. And in my experience many engineers are worse at this than they think.
Germany is kind of a developing country regarding software. There are good companies and many not so good ones. The things you are talking about, that comes after being able to write code, good code even. Architecture is not going to save you if you cannot write 200 lines of good code. Now, Leetcode doesn't test for good code but for "working" code AFAIU. And I'll be the first to say that code is written for humans to understand, but part of that is problem-solving ability.
Apart from that, I'm just passing on what I have heard from people. It's not first hand, but I believe it.
Apart from that, I'm just passing on what I have heard from people. It's not first hand, but I believe it.
So tired of seeing this guy. LinkedIns algorithm has decided my feed needs to be incessantly bombarded with his marketing rather than the updates from my connections that I actually want LinkedIn for. Is there any way to turn off non followed people entirely on my feed?
The title is highly misleading. The author of the article wrote a tool to "cheat" in LeetCode interviews, violated numerous interview policies, and stepped on a lot of toes. All while trying to pitch their product as both a company, and a moral imperative to get rid of leetcode.
This monster took advantage of trillion dollar companies?!
So he invented a program to do complicated work.
Agreed. This reminds me of the kids doing all kinds of dangerous, immoral stuff on TikTok or whatever just for views.
He is a legend. LeetCode is modern day slavery.
It's the best way to interview candidates by far.
What would you prefer instead? X years in stech stack, keeping devs boxed in their role for their entire careers.
What would you prefer instead? X years in stech stack, keeping devs boxed in their role for their entire careers.
I prefer interviews that determine what a candidate can do. The best so far was to hand them a laptop with a work ticket and some IDEs pointing to code that needed fixing.
Monkeying some algos allows devs who can go leetcode algos and are crap at everything else. I've seen their work after hiring.
no i would prefer code review. that's how i hire in my companies and we are more productive than faang.
is the person that made you take those interviews in the thread with us?
Oh please, spare the drama.
Average of 6 people involved in an interview loop. Let's say they make $250k a year each fully loaded on average (at the FAANGs, I'm probably underestimating). There's a 30-minute prep, a 30-minute debrief, the actual hour long interview, and probably 15 minutes of writing up the notes about the interview. Let's call it 2 hours. Taking the rough 2000 hours a year, that $250k is $125 an hour. So $250 per person times 6 people is $1,500. Times 4 companies is $6,000. Now, trade that 10 hours back for tickets. Who knows what the tickets would be, but it's 10 hours of work, with some likely positive ROI. I won't pretend to put a value on that, but it's non-zero.
He openly admits that he had no intention of taking the offers for these companies. He just wanted to test his cheating tool to see if it would work.
If that doesn't scream theft to you, I don't know what would. And sure, can FAANG pay the bill? Yes. But should they have to? What if he did it at a tiny 50-person bootstrapped startup? Can he be called a thief then?