Why I think the tech interview process is broken

I have had more than 5 interviews throughout the past two weeks. Some were on Skype, some on the phone, some in person. Almost all 40 of my classmates in the computer science master’s program had as many interviews and we discussed how they went. Here are some questions I was asked or I heard people were asked:

1. How do you find the log base 2 of a number?

2. Given this array of integers, what is the optimal way of doing X? What is the time complexity?

3. Given this board game, how do you find the optimal game? What are some edge cases you need to account for?

4. What is the difference between a tuple and a list in Python?

5. What is the formula for calculating X?

There is a lot to say about each one of these questions and I’m not going to cover each pitfall. There is one thing that they have in common: they are not near optimal in terms of the information they convey to the interviewer on my skills, my knowledge or how my brain works. All they are is a bunch of silly puzzles some people at Google et al. invented decades ago, because they thought they were less controversial than IQ tests, which were proved to be entirely pointless, at determining how fast the new hire would produce code that brings the most profits. As a bonus, you would learn if the interviewee knows how to use data structures and how not to crash the company’s servers. If this is what the almighty Google is relying on to build their multi-billion dollar business, who are we to question their practices? Did I mention almost 95% of Google’s revenues come from their search advertising monopoly?

People have been obsessing over this book, Cracking the Coding Interview. There are people who study this book and the likes for months to prepare for whiteboard interviews. All so that they can wear that Google t-shirt, feel proud and make their families back home proud. There is something very obvious that perhaps even the author herself is missing: the interview process can be cracked and cannot be relied on to drive technical innovation. The process merely becomes an extension of the broken education system. Do a little bit better than your peers in high school to get As. Do extracurriculars. Study these books to get good scores on SATs. At whatever cost, be admitted to the best universities preferably with scholarship. Again, perform a little bit better than your peers in university to get As. Win hackathons. Study this book to make a good impression during the interview. At whatever cost, be offered a position at the company preferably with a 6-figure salary so that you can pay your college debts. The same story that you thought you were getting away from.

Harvard rejected Jack Ma for its MBA program. Peter Thiel didn’t get the Supreme Court clerkship. Thiel says, “there often comes this tremendous price that you stop asking some bigger questions about what’s truly important and truly valuable. Don’t always go through the tiny little door that everyone’s trying to rush through, maybe go around the corner and go through the vast gate that nobody is taking.” Why is the tech interview process trying to rush hundreds of thousands of brilliant candidates through this tiny little door? Why should we pay this tremendous price of employers’ misconceptions about the dimensionality of a candidate’s mind?

Employers tell themselves that it’s not the answer the candidate gives, but it’s their approach. Let’s not kid ourselves. The candidate’s reasoning process on these questions is rarely ever representative of their reasoning process on much more realistic and complex problems in general. Not to mention how the artificial stress imposed on the candidate in front of the whiteboard is reminiscent of old school teachers’ oral exams. The same way no one cared if you had an imaginative approach in high school or college to give you an A, most employers don’t care if there is something unique about you that might give them an upper edge in their fight for market dominance and the innovation it requires, hence the tremendous price they pay.

I am going to resist and not memorize anything I can get my hands on in less than 10 seconds with a Google search or a textbook lookup. I am going to resist and try to see more in computer science than a bunch of algorithms and formulas. I am going to try and see this as a form of art. I am going to prove where things can go once we stop evaluating people based on how fast they can code, how careful they are or how much they can dump in their memory.

This industry needs change.

Poems, everybody! We need poems, no less.