Interview with Penske Media's Data Scientist – Andrew Maguire

Published January 23, 2017   |   

Andrew has been working in Analytics/Data Science for 7 years in various roles across many different industries. He is a Data Scientist at Penske Media Corporation focusing on both data engineering infrastructure as well as applied business analytics. Prior to this position, he worked at Google (marketing analytics, then local data quality), Accenture’s Analytics Innovation Centre (consultancy), and Aon’s Center for Innovation and Analytics (product development team).
This article is an excerpt of Springboard’s Guide to Data Science Interviews.
What do you look for when you’re hiring candidates?
Beyond meeting the basic requirements from a technical and experience point of view, I’d say enthusiasm, willingness and ability to continually learn new things are key..
A good attitude is super important, so someone who is able to also tell me about their weaknesses as well as strengths is a good way to draw this out (sometimes selling too hard is a bit off putting; humility is much better).
Being approachable, open and honest is something that’s key on the ‘team fit’ side. You don’t have to know the answer for everything but being able to work with others to come up with a decent solution is crucial.
What’s the best piece of advice you can give to people going through the data science interview process?
On the technical stuff, take your time, write stuff down, and ask clarifying questions. Also don’t be afraid to tell them if it’s an area you’ve not worked on before or an algorithm you’re not that familiar with. Being able to admit when your knowledge is limited is super important as a data scientist; continually learning is one of the most important skills required.
Make sure you have two or three data science ‘stories’ you can chat about with an interviewer that touch on problem formulation, data wrangling, analysis and insights, visualization and stakeholder communication. Try to get the balance right between cool nerdy technical stuff and showing business understanding and insights. These ‘stories’ can be projects from your previous roles, college assignments, or projects you did on your own time. Get good at spotting openings from interviewer questions to use your stories to show concrete examples and experience. I find that chatting (in detail) about projects the candidate has done in the past is the best way to get a proper feel for them (and best place to probe deeper from), so make sure you make it easy for the interviewer to be interested and excited to ask you about some project’s or example’s from your CV.
What kind of interview questions do you like to ask? What are you trying to test?
What’s the biggest or most complex dataset you have ever worked with? What problems did it create? (Trying to begin a discussion here that can lead into judging data wrangling skills and experience)
Give me an example of a time when you analysed a dataset, and communicate your findings back to the business. What was the problem faced? What did you find? How did this affect the business? (Touch on the extracting business insights and communicating back to stakeholders aspects)
I ask questions very related to what is on the CV, so if it’s a project from a previous role for example, I want you to explain what the problem was, what sort of data you used, how you used it, what the insights were, and how this all fits into the wider business. Choose what you put on your CV very deliberately. If you find it hard to get all on two pages then maybe have different ‘types’ of cv’s you might use for different types of roles.
Finally, I ask candidates to give me an example of a time when they failed, then add what they think went wrong and what they would do differently in future. This is something that comes out of HR 101, but I like to hear what they have to say 🙂
What is different about how Google hires data scientists from the rest of the industry?
I’m not sure there is too much of a difference anymore. Generally it depends on the specific role. For very specialized positions that are often more like research or fellowship positions, you would get much more detailed technical questions and problems to drive into the relevant area of expertise in very fine detail. For more generalist or business related roles, the focus is more on the right mix of technical skills, business understanding, working in teams, and communicating results to stakeholders.
The main difference in Google is that you have a lot more interviews and meet more people, so behind the scenes there are around 6+ people who have all met you and probed you from their own different angles. These people all have a different view of you and your strengths and weaknesses and must come to a decision and conclusion together that typically involves trade offs. Being able to show decent level of competence across the board as opposed to being a rockstar in one area but letting yourself down in others will generally serve you well. This is where attitude and being easy to get along with can be most important; even if you fall a little short on one of the competencies, if they like you and feel you could easily get up to speed in that area in a few months, then it’s less likely to be a deal breaker.
Want more advice on how to ace the data science interview and become a data scientist? Check out this list of 109 commonly asked data science interview questions, with solutions and discussion.