I Tested Startup Developer Candidates with Baekjoon 1000
Whenever I hire developers for a startup, I always agonize over how to structure the technical interview.
Recently, I had a junior candidate solve Baekjoon 1000 (basic arithmetic) and a simple array problem, and the results were shocking. People whose resumes boasted Baekjoon Platinum and CodeTree Gold were unexpectedly struggling with basic syntax. Of course, some of them might have just been nervous.
So I decided to change my interview approach. Instead of high-difficulty algorithm problems, I had them build a CRUD feature that you'd actually see in a real service, and watched how they refactored through debugging scenarios. That worked much better.
What I learned as an interviewer is that once you reach a true senior level, debugging ability and a systems perspective matter far more than algorithm scores. Instead of just grinding through YouTube algorithm-recommended problems, it seems more helpful to dig into lots of real-world repositories.
8 answers
Agreed. Basics are just as important as algorithms.
Yeah, that's right lol. I'm Platinum and I see a ton of people still struggling with basic grammar.
Well, Baekjoon 1000 is so easy that, ironically, if you get nervous you might end up making a mistake.
I've also seen people who actually hit Platinum tier struggle with pointers during interviews. Seems like they were only used to solving problems. I think it's true that practical skills matter more.
I agree with the advice to read lots of repositories. But expecting CRUD from a newbie is a bit... I think giving assignments is better than Baekjoon problems.
Wait, if you're Platinum and can't even handle basic syntax, doesn't that mean you just half-heartedly solved the algorithm problems? If you had real skill, problem 1000 wouldn't be a problem at all. Seems like a weird case.
What's the sample size? And you're generalizing from that?
Especially for startups, hands-on code maintenance skills are crucial. I think it's much better to filter candidates with algorithm coding tests and then, for those who pass, evaluate things like refactoring or error handling. Changing the interview format was a good call.