The "Data Scientist" Interview: Where to begin?

Vyas Anirudh
Apr 23, 2023
5 min read

The entire process of searching, preparing and nailing that data science interview is definitely a daunting endeavor. There are many different variations in the interview process depending on the company you are applying for and their job requirements. There is also a lot of mental stress in trying to deal with rejections and moving on to the next interview.

But fear not, for I am going to break down the data scientist interview process into few easy to understand digestible steps based on what I have read and experienced myself. I hope that this article will point you in the right direction and help you on your mission to get the job you desire. I will also be providing links to many different resources when necessary. Let's begin!

Applying for the job: Don't put all your Eggs in one Basket 🧺

This is where the journey begins, by sending your portfolio to the company for the particular job that you are interested in applying. There are multiple channels through which you should be applying for the job. This is important as it increases your chances of getting picked for an interview. The following are a few of the common ways to do it:

Job boards/Job sites: Websites such as LinkedIn, Glassdoor, Indeed, Robert Half and many more are probably one of the first things that you look into to start applying.
Company Websites: This is another good source to find the relevant jobs since it could be the case that the company does not use the above mentioned job boards/sites to post the job but instead puts them in the "Careers" or "Join us" section at the very bottom of their webpage
Cold E-mail: This is equivalent to throwing a dart while blindfolded and hoping that it hits the bulls-eye. With this approach, you would first need to find out the recruiter's e-mail either through company hosted events, workshops or even through their LinkedIn profile. Next, you would be sending out an e-mail to them talking about how you came across their profile and what makes you the ideal candidate for the job. I will not go into the details as that is a whole new topic for another day. You can refer to this book for a good overview with practical examples on this topic.
Coffee Chats: "Coffee" here doesn't necessarily mean you buy them a coffee unless it is an in-person meeting. This is where you send out an email or a message to the recruiter, senior developer or product manager on the team requesting to have a quick 15-minute chat about the role and also getting to know the other person's job responsibilities as well. Just remember one thing and this is very important, do not straight-out ask them to get you the job. They cannot do this. You will lose your connection with them and also might get a bad impression from the team if later you do decide to apply through their company website and it turns out the person you had the coffee chat with is the person reviewing your profile.

The Coding Round: A heavy brain-workout 🧠💪

Assuming that you have been picked for the interview, one of the most common first interview rounds is a coding test(depending on the job requirements). This is where you will be tested on your data structures & algorithms, programming skills and your ability to solve math-oriented problems using a programming language of your choice. You will be given around 2-3 questions and then be asked to provide a solution within a given time frame(usually 60 - 90 minutes). As an example, you can be asked to come up with an algorithm to find out if there are any duplicates in the given list of people's names.

Python, R, C++ and Java are the most commonly used programming languages when it comes to Machine Learning. Try to stick to the one that you are most comfortable with, since solving the problem is more important the language used to solve it. To help you get an idea on the types of problems you might face in this round, I have listed a few sources:

LeetCode
HackerRank (This is my personal favorite 😎)
CodeChef
Codecademy(more useful for prep)

Another variation in this round would be a live-coding interview where a developer from the team would usually be on the call with you while you are solving the questions and will be asking you to provide an explanation for your solution as well so be prepared!

Phone Screening 📞

This will be a 15 - 30 minute phone call with the recruiter. The purpose of the call is to get to know you better and check to see if you are who you say you are on your resume. Typical questions will be:

Tell me about yourself
Why do you want to apply for this job?
Why do you want to move from your current role?
Can you talk about a project that you did recently?
What are you salary expectations?(This is a good article for reference)
Are you willing to relocate for the job?

This part of the interview may usually come at the very beginning even before The Coding Round.

Take Home Assignment ⭐: A homework basically

You have gotten this far which is really great! But the story hasn't reached the climax yet. In this part you will be given a document stating some kind of a real-world data science problem. You will be given the problem statement, data, and at times some starter code in the form of scripts or a Jupyter Notebook and expectations from your answer.

You are expected to provide a a short presentation/analysis report on your solution and might also be asked to give a presentation to a Senior Developer/Product Manager on the team as well. You will be given enough time, usually from 2 days to a week at most, and there will be somebody to assist you in case you get stuck in terms of the logistics and understanding the question itself.

If this sounds very new to you then I would highly recommend getting yourself familiar with how data should be analyzed, cleaned and presented with insights. Mastering every machine learning algorithm will only teach you about the algorithms themselves. But in reality you would also need to deal with a lot of messy data and turn in into useful business insights. Kaggle and DataCamp can definitely help you get started.

The whole point of this round is to test your "Data Story Telling" capability. This might sound very fancy at first but you can refer to this course to help you understand it better.

Interview with the team 👩💼👨💼

This particular round might take place even before the Take Home Assignment or it might be the first round that you face. Here, you are invited to a 60 - 90 minute meeting with a few members from the team which might also include your hiring manager. You will be asked a variety of questions ranging from technical to behavioral.

Technical Questions: This is where you need to showcase all the effort that went into your preparation. The questions will be focused on machine learning algorithms, the underlying principles and methods of dealing with different types and sizes of datasets etc. Depending on the job responsibilities you might also be asked about your experience with various cloud technologies and deployment.

Behavioral Questions: Most of the questions here will be similar to the Phone Screening interview with the difference being that you are also answering your potential manager/VP of the team. The only advice I can give here is to be honest in your answers.

There might be other types of rounds as well but the above mentioned aspects broadly covers the full data science interview process. Hopefully you now have a good understanding and are better prepared to face the interview confidently 😀

One thing to remember is to not contemplate the past(interview rejections) for too long. Just learn and move forward!