Traditional technical interviews are based on a man-made situation that is different from the real world

It's fundamentally problematic, not just intellectual questions themselves, bringing applicants into a room, overriding technical issues on the spot without any necessary resources, or writing code on a whiteboard .

Two years ago, I wrote an article "Why Newcomers Do Not Coding," to express my disdain for the industry-wide interview process of software engineers, mentioning in particular the kind of intelligence questions and binary searches like Microsoft and Google problem. Look, this week Google's boss of human resources frankly said: "The problem of IQ is a waste of time."

Traditional technical interviews are based on a man-made situation that is different from the real world. It's fundamentally problematic, not just intellectual questions themselves, bringing applicants into a room, overriding technical issues on the spot without any necessary resources, or writing code on a whiteboard.

There are quite a few people who share the same viewpoints as me. Many of them have proposed improvements, which are basically the same, and can be summarized as follows.

~ ~ With five or six basic technical problems quickly filter out the technology, but the people. Ideally, this step can be done by phone and the number of people being filtered out will be more than enough to surprise you.

~ ~ Face to face with candidates (not local also by phone) to exchange what they have encountered technical problems, tools used, made decisions and so on. Google's data proves that this "behavioral interview" is the most useful. But when discussing technical concepts, do not torture each other with questions that must be accurately answered.

~ Most importantly, discuss what the candidate did before. Or if the other party has a code on Github, let him talk about it.

~ ~ Try to establish a "cultural fit." Note that here only refers to the work culture. Some people just do not like you, do not subconsciously turn them out. Do not make the company a brotherhood.

~ ~ Finally, if you have reached this step, then arrange a "pilot project." Find a real-life project that's fit, self-contained, non-critical, but will be released when it's successful, pay for candidates to do the last week and pay close attention to the process.

The next question is with and without.

In doing so, employers can learn more, waste less time, and can even clean up small, real-world projects that are always on the shelf. The only downside is that this approach is poorly scalable and may not be suitable for large companies.

Well, I'm a bit headline party, technical interview did not die. The whole purpose of the interview is to understand the true level and efficiency, since now can be done, why not?

Jon Evan is a Canadian novelist, journalist and programmer. Possession of electronic engineering background and more than 10 years of software development experience, works have been translated into multi-language publishing. Representatives have "Dark Places", "Beasts of New York" and so on.