Algorithms interviews: theory vs. practice

#artificialintelligence 

Real is in quotes because I've passed a number of interviews for reasons outside of the interview process. Maybe I had a very strong internal recommendation that could override my interview performance, maybe someone read my blog and assumed that I can do reasonable work based on my writing, or maybe someone read some of my open source code and judged me on that instead of a whiteboard coding question (and as far as I know, that last one has only happened once or twice). The reason it's arguably zero is that the only software interview where I inarguably got a "real" interview and was coming in cold was at Google, but that only happened because the interviews that were assigned interviewed me for the wrong ladder -- I was interviewing for a hardware position, but I was being interviewed by software folks, so I got what was basically a standard software interview except that one interviewer asked me some questions about state machine and cache coherence (or something like that). After they realized that they'd interviewed me for the wrong ladder, I had a follow-up phone interview from a hardware engineer to make sure I wasn't totally faking having worked at a hardware startup from 2005 to 2013. It's possible that I failed the software part of the interview and was basically hired on the strength of the follow-up phone screen.