Data Science Interview Mistakes – by Roopam

Are you preparing for data science interviews? To help you in your preparation, in this article, I will discuss some of the worst mistakes candidates make in data science interviews. In the previous article, we discussed the structure of data science interviews along with sample questions and preparation strategy. Many times I have noticed that well-prepared candidates fail to clear interviews. I will highlight the reasons I strongly believe are at the core of failure for these candidates. On some level, the root cause for this can be traced to selfie mania.

Selfie Mania

A selfie is a self-portrait photograph usually clicked through a smartphone. The popularity of selfies can be assessed by the fact that on Instagram alone 259 million posts are tagged as #selfie. Selfies feed well into the human ego and self-centered universe. No wonder the planet is gripped with selfie mania. Obsession with selfies can also have fatal consequences. In 2015, selfies caused 27 deaths worldwide. These victims of selfies in their zeal for a perfect picture of themselves lost touch with reality around them and made fatal mistakes. I guess, capturing the moment and living in the moment are two different things.

The lack of awareness of surroundings is also at the core of failure of candidates in data science interviews.

7 Mistakes in Data Science Interviews

This lack of awareness is manifested through these 7 worst mistakes in data science interviews.

1. Not Engaging in a Conversation
2. Not Directing the Course of the Interview
3. Not Using “I don’t know” Judiciously
4. Not Taking the Opportunity to go into Details
5. Focusing on Answer Rather than Approach
6. Not Asking Questions
7. Taking Failure Personally

I will discuss these mistakes in detail in the subsequent sections and also suggest ways to avoid them.

1. Not Engaging in a Conversation

Almost all interviewers want you to succeed in the interview. At the same time, they are also scared of making a bad decision in choosing their future colleague. Resolving this conflict for the interviewer is essentially your goal during the interview towards your success. Nothing is a better tool to resolve this interviewers’ dilemma than a good conversation.

The art of conversation is the art of hearing as well as of being heard. – William Hazlitt

Now if you could engage your interviewer in a free-flowing conversation you will mostly achieve your goal. Conversation is a two-way street. Let me share a valuable secret I have learned in all my years in data science consulting. A client presentation is almost always a success when your client gets engaged in your results, and speaks more than you. This golden rule also applies to an interview. You want to give your interviewer enough opportunity to speak, and being heard. This requires you to have a good awareness of your surrounding and avoid selfie mania.

Moreover, in any interview, you will always speak more than the interviewer. A good ratio for your vs the interviewer’s speaking time is roughly 70% / 30%. Since you are going to speak more than the interviewer the onus of directing the course of the interview is also with you. This is also the second worst mistake in data science interviews.

2. Not Directing the Course of the Interview

A few years ago I had interviewed a candidate who blabbered about artificial neural networks (ANN) for 20 minutes during the interview. After that, he acknowledged that he did not understand ANN very well as he had always worked on logistic regression. To make things worse, he was the one who started the conversation on ANN by highlighting it in his CV. He, of course, could not clear that interview.

I can’t change the direction of the wind, but I can adjust my sails to always reach my destination. – Jimmy Dean

The elements this candidate missed are found in drama and theater. As part of my hobbies, I had a short stint in professional theater. I was not a very good actor. Since it was all fun and games it never hurt my prospects in my career. But I learned a great lesson from the theater that still helps me in my data science consulting. I was told that I sucked in acting because I didn’t follow my cues very well. By the way, cues are anything said or done, on or off stage, that are followed by a specific line or action (source dictionary.com).

What you write in you CV or speak during the interview is a cue for the interviewer to progress in the interview. Don’t talk about ANN if you don’t know enough about them. During the interview, it is your responsibility to direct the interviewer towards concepts you are comfortable with.

A desire to be in charge of our own lives, a need for control, is born in each of us. It is essential to our mental health, and our success, that we take control. – Robert Foster Bennett

3. Not Using “I don’t know” Judiciously

While directing the course of the interview, there are going to be situations where you will find questions you have no clue how to answer. It is judicious to acknowledge your ignorance followed by a statement like “let me read a bit about it and address this next time we meet”.

Teach thy tongue to say ‘I do not know’, and thou shalt progress. – Maimonides

4. Not Taking the Opportunity to Go into Details

Some time ago I interviewed another candidate who was doing quite well in the interview till I asked him to explain one of his projects about price optimization in detail. Initially, he brushed off the details. I gave him enough cues to explain his points in detail but he simply didn’t relent. At one point I almost pleaded for him to use the whiteboard to explain his thoughts in a structured way. The guy refused to touch the whiteboard marker, and hence never hit the mark in this interview.

Beware of the person who can’t be bothered by details. – William Feather

For data science interviews, always go well prepared with at least one of your favorite data science projects. Make sure you know every small detail about this project. You need to know how data was procured, prepared, and analyzed. Moreover, your understanding of the business problem and the way your approach solved the problem will help the interviewer choose you as his future colleague.

Attention to detail is of utmost importance when you want to look good. – Carolina Herrera

5. Focus on Answer Rather than Approach

Success is a journey, not a destination. The doing is often more important than the outcome. – Arthur Ashe

Douglas Adams in his funny book trilogy The Hitchhiker’s Guide to the Galaxy described a super smart computer called Deep-thought. Deep-thought was asked the answer to The Great Question of the Life, the Universe, and Everything. Deep-thought engages in its calculation to this grand question. After 7.5 million years of effort, Deep-thought finally revealed the answer to this ultimate question as ‘forty-two’. Deep-thought of course never explained the reason for his answer.

This is a funny story for everyone except interviewers. Interviewers always care more about your approach to answering the question than your answer. Please don’t torture your interviewers by acting like Deep-thought.

Character is a journey, not a destination. – William J. Clinton

6. Not Asking Questions

In school, we’re rewarded for having the answer, not for asking a good question. – Richard Saul Wurman

Looking for answers but not the questions is something ingrained in us through long years of schooling. However, as a data science consultant, you are expected to ask questions as your foremost task.

My greatest strength as a consultant is to be ignorant and ask a few questions. – Peter Drucker

Similarly, in data science interviews, every good interviewer will provide you with an opportunity to ask questions at the end of the interview. This is your opportunity to assess whether this new job is a good fit for you. Believe it or not, the questions you asked at the end of the interview also help the interviewer to assess your seriousness and maturity required for the job.

He who asks a question is a fool for five minutes; he who does not ask a question remains a fool forever. – Chinese proverb

7. Taking Failure Personally

Finally, some of you may have to go through many data science interviews before securing a job. Moreover, sometimes even after a great interview, you might not secure an offer. Taking these failures personally is the worst thing you could do to hurt your chances of securing a job offer soon. There are several reasons, beyond your interview, that could be responsible for you not securing an offer. Sometimes there are other better candidates. Other times the position you were interviewed for has been kept on hold because of financial pressure on the organization. If you will fall for selfie mania then you will mostly internalize the failure without seeing it in the right context i.e. the complete picture with it’s surrounding.

A failure is a man who has blundered but is not able to cash in the experience. – Elbert Hubbard

The first thing to do after an interview is to rationally identify things that you did right in the interview. Also, recognize uncomfortable moments during the interview and their reasons. The idea is to eliminate these reasons before your next interview.

Success is not final, failure is not fatal: it is the courage to continue that counts. – Winston Churchill

Sign-off Note

Have a look at this picture that has gone viral on the internet. After all, capturing the moment and living in the moment are not the same. All the best for your data science interviews.