This story seems like a bad example of a false negative. There were claims by Google insiders that Max was given a particularly easy interview as a formality. I am by no means a brilliant programmer and also rusty with algorithms, but given the structure and question, I would be surprised if most people couldn't figure out how to invert a binary tree within a matter of minutes.
Max himself later opened up[0] and admitted to being difficult to work with. It is entirely likely that he was rejected based on his personality and not his ability. As someone who contributed to Homebrew many years ago I would not be surprised if this was the case. In their own words: "I am often a dick, I am often difficult, I often don’t know computer science". I am not sure why any company would want to hire someone like that and put the culture of the team in jeopardy.
I love your response because I have a PhD in Computer Science. I do LeetCode as and CodeWars problems in the mornings for fun (I like coding and helps me avoid getting rusty) and I have been a hiring manager for over 10 years.
But yeah, other than the "sexiness" of CS concepts it's only because of the "narrative" we want to push.
So you do think Leetcode is useful ? What's up with the flex ? Congratulations on the PhD and job, I guess ?. Looks like you could definitely invert a binary tree.
Leetcode is good for some things, but in my experience is it not a good factor to filter Software Engineers (it is a good metric to filter candidates to form a team for the IOI or similar).
This is a response I wrote elsewhere:
I've been growing Engineering teams for the last 10 years as hiring managers in different startups. At some point in our startup we had those kind of HackerRank questions as filters.
The thing we realized is that those sort of interviews optimize to hire a specific type of very Jr Engineers who have recently graduated or are graduating from CS. That is because those are the people that have the time to churn these types of "puzzle" problems. Particularly, there are 3 types of recent graduates from CS or related fields: The ones that don't know crap, the ones that focus on these sort of problems, and the ones that are "generalists" because they dove into all sort of subjects during their degree.
I found out that the Jr people that excel at those sort of problems have a huge learning curve to climb to be productive in "production", real life environment. On the contrary, the "generalists" work better.
We stopped doing those sort of algorithm puzzles interviews after that realization, and we started getting really good Engineers with great real-life experience.
Yeah, nothing wrong with not having memorized how to do that - it's certainly not often useful in practice. But if you're given the definitions and a couple minutes to figure it out (which, in an interview, you are), and you actually can't come up with the 10 lines of code to do it, maybe that question is working as intended after all...
> But if you're given the definitions and a couple minutes to figure it out (which, in an interview, you are), and you actually can't come up with the 10 lines of code to do it, maybe that question is working as intended after all...
or maybe the person struggles to work things out in an extremely stressful situation
I speak as somebody that finds interviews Extremely stressful and if working there is regularly my interview levels of stress then please do reject me.
https://twitter.com/mxcl/status/608682016205344768?lang=en