How to Conduct an Algorithms Interview

Jay Wengrow’s article “How to Conduct an Algorithms Interview” leads off the August issue of PragPub. It starts out like this:

By the time you read this article, my book A Common-Sense Guide to Data Structures and Algorithms will either be in print or be going to print shortly. While the book aims to demonstrate the power these concepts can provide in everyday programming, I suspect that a fair percentage of people who buy the book will do so to help prepare for technical programming interviews. (I even list that as one of the book’s benefits in the introduction.)

Yet, there’s an irony here that simply cannot be ignored: I just authored a book on algorithms, yet I’m certain I’d fail plenty of whiteboarding and algorithm-based interviews. How can this be? If someone understands data structures and algorithms, how could such a person fail an interview that quizzes that person on these topics?

This reveals the big flaw with our interviewing system: There is not one, but two distinct skillsets at play here. The first skillset is the ability to reason about algorithms. The second skillset is the ability to succeed at an algorithm interview. And while these two skillsets may have a little overlap, they are actually very different skillsets.

The ability to reason about algorithms is the ability to understand an algorithm’s efficiency in terms of time and space — that is, how fast does it run and how much memory does it consume. The ability to succeed at algorithm interviews, however, involves many other things, including prior knowledge of certain algorithms and data structures that may have no practical application other than on technical interviews.

I recently showed someone the various books on the market that have titles along the lines of, “How To Succeed At Programming Interviews,” and his response was on the mark. He said, “Shouldn’t the book be called, ‘How To Program?’” From the fact that there are specific books on how to succeed at programming interviews, that clearly reveals that interviewing to become a programmer has little correlation to what programmers actually do.

There’s been a lot of debate on this topic over the years. Proponents of algorithm-based interviews believe that such interviews can reveal much about the candidate’s thought process, as well as reveal their aptitude for writing efficient code. I can appreciate that, but I believe that such an interview has to be conducted correctly to properly achieve those results….

Jay Wengrow is the founder and CEO of Actualize, and the author of A Common-Sense Guide to Data Structures and Algorithms.

The August issue of PragPub is on sale now at The Prose Garden. Two measly bucks. Twenty for an entire year.

Comments are closed.