Hacker News new | ask | show | jobs
by jsgoller1 2737 days ago
Interviewing. I spent about 3-4 months this year working on my interviewing and competitive programming skills via LeetCode, reading The Algorithm Design Manual, and taking classes at Bradfield School of CS in San Francisco. I feel like it's made a world of difference. I know Hacker News tends to hate on programming interviews, but it really pays to be good at them (and trust me, it's a learnable skill). I do agree that the algorithms-and-whiteboard method can be highly problematic, but I've come to firmly believe that the time for criticism is when you're designing the interview for a candidate, not being interviewed.
2 comments

Agreed! It's definitely annoying to train something which doesn't seem particularly related to our jobs, but it's worth the effort.

The career impact of becoming a better programmer or leader is tiny compared to the impact of moving to a better company.

If you're a free agent and you're not interested in interviewing or working for big companies, then I would say the top skills to learn are listening/negotiation and maybe time management, unless you're already great at those things. Same advice if you're already at one of the best companies and don't need to interview :)

You spent 3-4 months honing your interview skills, that sounds very expensive.

Are you claiming that this practice made you a better programmer overall or just better at interviews?

It isn’t just hackernews hating on the current way of doing programming interviews, it’s a widespread sentiment in the industry. The current way of doing interviews seems both excessive (requires 2 months of practice) and easily games (with 2 months of practice...). Who is winning from this?

> You spent 3-4 months honing your interview skills, that sounds very expensive.

Assuming a 25% pay bump, it more or less pays for itself in a year.

Not counting the effect of compounding. Bumps are typically by percent, so every point you chisel out early in your career is valuable.

Assuming you couldn’t get that job any other way, sure. However, even most Googlers would think that is excessive (even if they did it themselves, most of them would not admit to it).
The current way of doing interviews seems both excessive (requires 2 months of practice) and easily games (with 2 months of practice...). Who is winning from this?

While I have never had a leetCode style interview, I’ve also far removed from Silicon Valley. That being said, if that’s what my local market required, that’s what I would do.

“The [job] market can stay irrational longer than you can stay solvent”

Both - my interviewing skills got better mostly from doing mock interviews and exposure to the questions that would be asked during an interview, but I'd say my programming skills got better from learning more about algorithm and data structure design, as well as carefully thinking about edge cases so as to beat the battery of test cases that each LeetCode problem has.
We really have to be careful here, because many CS programs have or are planning CS courses specifically designed around LeetCode problems as a way of giving their students a leg up in the interview process. If these kinds of interviews become more widespread (and there are a lot of companies trying to copy google here), this will only intensify and possibly begin push out other CS courses with more useful content.

That kind of future would be very dystopian for the field.

From a very pragmatic standpoint, the most useful courses in college are those that help you get a job.
Taken to an extreme that can get ridiculous (eg if colleges start teaching courses titled “fake it until you can make it”).
What’s the difference between “faking it until you make it” and “learning on the job?” I don’t think I have had a single job in 20+ years where I didn’t spend the first six months in Imposter Syndrome territory.
I'd also argue it's more like a 2x pay jump if you're early career. I'd love to know about companies who paid that well and didn't require leetcode