Modern tracking and analytical tools allow us to collect an unprecedented amount of data about every visitor of our website, or every customer who walks into our store.
We can track everything including the effectiveness of employees, processes, and investments. Data means power, without a doubt, and the success stories of big IT giants all around the world just confirm the assumption.
Having terabytes of raw data, however, means nothing–unless we can interpret them, and draw conclusions and predictions for the future. Unless we can use them practically.
That will be your role as a data scientist. To gather raw data, to process and analyze them, to design predictive models and algorithms, and to propose solutions to challenges and ideas on improving the business process.
In this article we will look at some questions you may face while interviewing for this interesting job, including personal, behavioral and technical questions. We will also show you how to answer some of the questions. Enjoy!
Please tell us something about your experience (walk us through your resume).
Focus on the most relevant things. If you just graduated, talk about subjects and courses that helped you get ready for this role–programming, software testing, time row analysis, statistics, and so on.
If you have some experience under your belt already, speak about things you did in your former jobs, related to data science. This can be anything from analyzing data to making predictions of any kind, or using your programming skills to address any sorts of problems and challenges.
Try to speak with enthusiasm about your studies and jobs you had. The interviewers should get a feeling that you enjoy doing your work, that you aren’t in only for an excellent salary or recognition of your peers.
Which programming languages are you most comfortable with?
I suggest you to mention all languages you’ve applied in your work, and to always add a practical application, or an algorithm or program you developed while using the language (or combination of more languages).
Python, SQL, and R have come in front recently as programming languages for data scientists, but a knowledge of Java or C++ (or anything else) will always come handy.
They say that “The more languages you know the more you are human“. In programming we can say that “The more languages you know the broader are your possibilities as a data scientist”.
And if you are just starting are are skilled in one or two languages only, ensure the interviewers that you are working on your skills, and want to learn more programming languages in the near future.
Describe your most successful and least successful project.
Now, the key is to demonstrate that you do not work for the sake of numbers only. Certainly you have passion for coding and find pleasure in doing the job, but that’s not enough for the employer.
The most successful project is the one in which you managed to save a significant amount of money for your employer, or perhaps one when you eventually came to a model/prediction that helped them to beat their competition or foresee a critical challenge on the horizon.
Similarly, the least successful project is the one in which you made some bad forecast, or drew wrong conclusions, and it cost your employer a lot of money.
Show them that you understand why they pay you $100K+ salary. They don’t pay you so much so you can play with numbers and algorithms. They pay you so much becasue they know how valuable your predictions can be.
One more thing–when you talk about the least successful project, you should ensure the interviewers that you learned your lesson, and will not repeat the same mistake again.
How do you imagine a typical day in work as a data scientist in our company?
It is essential that you read the job description carefully, and also try to understand the reality of your job. Actually one of the most common reasons why data scientists quit their jobs (often after a few months) are unrealistic expectations.
You hope to write smart machine learning algorithms and draw conclusions that can take the company to the next level, but eventually all you do in your entry level data science job is working with a limited set of data and preparing some basic charts for managers, so they can boast about something during board meetings…
Now, everyone has to start somewhere and you can’t hope to do the most interesting work right from the start. If you apply for a big corporation job, you will probably start with assisting senior data scientists (or other professionals), doing some simple analysis and helping with parts of code.
Let me repeat it: read the job description carefully, and do some research about your prospective employer. This will help you avoid unrealistic expectations, and a bad outcome at the end of your interview.
Tell us about a situation when you struggled to communicate something to one of your colleagues (marketing manager, sales manager, anyone without technical knowledge). How did you get your message over?
The key is to convince the hiring managers that you can explain technical things in a business language, that you can translate the results of your mathematical and programming work to actual recommendations managers can work with.
And this is exactly the sort of situation you should pick for your answer. You came to some math conclusions, and you explained them in a simple way to the managers (or to other decision makers) who later applied them while improving the business process.
Ensure your interviewers that you understand the importance of effective communication, and know that your work would be worthless if the managers did not understand how to apply your conclusions in their work. That’s the attitude they seek in a great job applicant.
What do you want to accomplish in this job?
Many excellent data scientists are like children (in good means). They love to play with numbers, they love to experiment and try all kinds of things. They do not do it to achieve any particular result–they simply enjoy the process.
And while it would be difficult to let your imagination roam and come up with innovative ideas if you lost your passion for “playing with numbers”, you should still talk about some tangible accomplishments for your employer. Think about their core business. What they are doing, what is decisive for their revenue and profit, what factors come into play, and which of them you can influence as a data scientist.
This can be anything from calculating right interest rates to designing an algorithm that will predict (or analyze) customer behavior in some moment of the purchasing process. Just think for a while about the value you can bring to their team as a data scientist.
If you manage to identify anything tangible you can bring to their business, and point it out as something you’d like to accomplish in your new job, you can be almost certain that they will hire you–becasue such a mindset is rare when we talk about programmers and IT specialists…
Some other questions you may get in your data scientist interview (mostly technical)
- Tell us how we could sort a large list of numbers?
- How do you test your code? What kind of tests do you write, and in which language?
- What are software patterns? Which patterns are you familiar with?
- How would you build a search engine for a very large collection of documents?
- How would you explain deep learning to a ten years old child?
- Explain exploding gradients.
- Here is a piece of code (in some programming language). We purposely made some mistakes in the code. You have five minutes to find the mistakes.
- What cross-validation technique would you use on a time series data set.
- How do you plan to deal with a crisis of motivation, that typically arrives in any technical job?
- Describe the main difference between Point Estimates and Confidence Interval?
- In any five minutes interval, there is a 10% probability that you will see at least one shooting star. How likely are you to see at least one shooting star in the time span of 45 minutes?
- Explain SVM algorithm.
- What do you do about missing values when you analyze some historical data?
- Do you have any questions?
* You can also download the full list of questions in a one page long PDF, and practice your interview answers anytime later:
Conclusion and final tips
Data scientists are in high demand, and you shouldn’t find it difficult to get a job in the field. As long as you manage to convince the interviewers of your honest motivation and technical skills, and do not present some unrealistic expectations (either about your working duties or about your salary), they should give you a chance to prove your skills in a trial period.
Remember that data scientists aren’t locked in their comfy offices all day long, talking to nobody.
Meetings with managers and presentations of your conclusions to people with little or no technical knowledge forms an integral part of the job. For this reason they will test more than just your technical skills in an interview. What you expect from the job, what you want to accomplish, if you can explain technical things in a language of common people, etc.
Whether technical or behavioral (scenario) and personal questions prevail depends mostly on the person who leads your interview, and their level of technical skills. HR Generalist or CEO will ask you totally different questions than a senior Software Tester or a Data Scientist.
I suggest you to prepare for all types of questions (both technical and non-technical). That way you should be ready to ace your interview, regardless of who leads it… We wish you good luck!
May also interest you:
- How to overcome interview nerves – Do not let stress to kill your chances in an interview.
- Salary negotiation tips – You can ask for a lot as a data scientist, and we will show you how to get the best possible offer.
- Work portfolio for an interview – Learn how to prepare a selection of your best works, and how to use it to show the interviewers the value you can bring to their team.