How to: Pass a Silicon Valley Software Engineering Interview

Incidentally, the sorting algorithm described here is a counting sort:<p><a href="http://en.wikipedia.org/wiki/Counting_sort" rel="nofollow">http://en.wikipedia.org/wiki/Counting_sort</a><p>When you're trying to get nice low asymptotic time bounds on things, it can really help to remember counting sort, radix sort, and tries. Counting and radix sort are O(n), and tries are a really handy data structure with the same worst-case asymptotic time as the expected amortized time of hash tables. They also support in-order key traversal and prefix/range queries.

Very good advice. This is the best suggestion: "As I've said before - the interview is very very honest. Its about you, the whiteboard, and what you can do."<p>In the interviews we do we include a written exercise you have to complete before you come in the door so that we can sit and talk about a problem you've been able to solve at your own pace - instead of seeing how well you think on your feet. Even then people try to come up with the fastest solution instead of something that represents them well.<p>Remember that the interviewer is always more impressed with a right answer than a fast answer.

<i>Also - forget technical religion.</i><p>Hah! When I interviewed at Google, I turned down the invite to the next round because it felt more like a cult than an engineering organization.<p><i>if Google stopped paying me tomorrow, I'd still come to work</i><p>Exactly.

An old but good article. Might be worth amending the title to reflect that this is from 2007.

<i>They do that stuff in their spare time - its not just a job, its what they do because they love it.</i><p>Is there any other professional career where it's common practice to expect candidates to practice their profession in their spare time? My CPA ex never came home and crunched numbers for fun. My doctor never mentioned diagnosing his neighbors to relax.<p>This seems like a self-deluded way to weed out people who have families, physical limitations or injuries that prompt them to get away from their computer, etc.

His grandmother knew radix sort. It's well known that, on sufficiently short keys and sufficiently many records, radix sort beats the (n)ln(n) sort algorithms such as heap sort and, thus, beats the Gleason bound. How? The Gleason bound is only for sorting by comparing pairs of keys, and radix sort doesn't do that. Maybe his grandmother had a job with punch cards: The old punch card sorter used radix sort. Of course Knuth covers radix sort in TACP. However I've always thought that his claim that radix sort was especially good for a pipelined processor was wrong.<p>The whole column and its approach and values are wrong. If Silicon Valley and Google are recruiting that way, and I suspect that they are, then GOOD because it means that competing with them should be easy. Basically the author just nearly totally fails to understand what's important in computing or what qualifications are important.<p>It's an old story: In a well run technology company, HR is absolutely, positively forbidden to engage job candidates in any meaningful sense whatsoever on threat of immediate reassignment to the toilet squad. All HR can do is push paper and keep records, smile, bring coffee, tea, soda, or donuts, make travel arrangements, etc.