Hacker News Reader

Comment by oneepic | Hacker News Reader

I disagree because:

1) you don't have to know everything, only the topics they happen to ask. You are rolling the dice, but part of the time you'll be lucky. Besides, after reviewing for a few months (also did 4 yr degree in CS) I felt really strong in the algo topics. That said, I started half of my recent Google interviews thinking "how the fuck do I solve this? is this where I fail?"

2) Aside from the topics, you can absolutely drill the process. Drill a basic flow like understand/clarify the question, do examples, mention a brute-force, etc. when you do LC problems.

View on HN

And for 1)--a good interviewer won't actually mark you down for not knowing some fact, or not being able to think of some trick. I've passed interviews where I've just said, "Hey, I don't know [that particular thing]. How should it work?" or "I'd Google the exact algorithm for X; I'll pretend I wrote that and call it here..." or similar.

Good interviewers want you to pass, and aren't just giving you a test of arbitrary programming trivia.

Technical interviews are about culling the incoming applicants on the assumption that passing many good hires is preferable to letting one bad hire through. So, IME, no matter how well intentioned an interviewer may be, they aren't ultimately looking for a reason to pass you - they are looking for any way to differentiate between candidates.

Lip service is certainly paid to 'everyone has to look things up', and doing a quick search won't necessarily count against you, but, IME, the hesitation and doubt that caused you to look things up will. With so little material with which to evaluate a candidate, absolutely everything that doesn't impress them is going to count against you.

> passing many good hires is preferable to letting one bad hire through.

This is certainly Google’s philosophy, but our industry doesn’t think with one mind on this.

And as for having so little material - to me this is a sign of a badly designed interview. Almost everyone is weak in one area or another. If your interview only assesses candidates in one way (eg via a coding challenge, or based on a single whiteboard problem) then you are making a decision with insufficient signal. Multifaceted interviews are good interviews.

Plenty of otherwise strong candidates are weak in at least one section of any assessment. And plenty of bad hires will still, for example, know trivia about data structures even though they don’t actually know how to program. Making a hiring decision based on a single metric leaves way too much to chance.

(Source: I’ve done over 400 technical interviews)

This is a lie at least as it corresponds to bigtech hiring today.

> "I'd Google the exact algorithm for X"

This is a fail in the interviewer's book. Read: "Could not produce an algorithm with less than O(n)" or "lacks familiarity with fundamental data structures"

I passed an Amazon interview by saying almost exactly that, so given Amazon is one of the FAANG companies? I can back up my statement.

Reply on news.ycombinator.com