Advice for Data Science Interviews
Check out an interview excerpt from Springboard’s Guide to Data Science Interviews. Determine how one can find data science interviews - and ace them!
By Will Kurt, Springboard.
This interview is an excerpt of Springboard’s Guide to Data Science Interviews which combines insights from recruiters, hiring managers, and successful interviewees into how one can find data science interviews-- and ace them.
What do you look for when you’re hiring candidates?
The biggest thing for me has always been a combination of creativity and genuine curiosity. In a startup environment new problems come up every day in a wide range of areas. One month you may be helping the product team add new features, the next helping sales improve their process, and still another helping marketing restructure their testing setup. The most valuable candidates are the ones interested in everybody in the company’s data related problems and always thinking of new and interesting ways to solve them.
What’s the best piece of advice you can give to people going through the data science interview process?
In my experience, all small companies and startups worth working for are really excited about the idea of adding a new data scientist to the team. They’re hoping your skills and experience will help them solve a range of problems they’ve been struggling with. Show up to the interview ready to listen to what they’re trying to solve and get them excited about solving things together. Every chance you get ask people what they’re working on, get them brainstorming with you about ways you could make their day better. Don’t worry about impressing people with your technical skills. There are thousands of candidates out there with amazing quantitative skills, candidates who really care and are excited are very rare. Leave the interview with everyone wanting to work with you on a project and they’ll be the ones hoping you say “yes”.
What kind of interview questions do you like to ask? What are you trying to test?
All I care about is how your mind works once it’s fixed itself on an interesting problem. At Kissmetrics I used to give out an open ended “homework” assignment. There was a really obvious approach to the problem (build a classifier), but I mentioned this and cautioned that part of the test was to see if you could come up with something interesting. The results of the assignment didn’t have to be long or complicated. What really mattered is that they started a conversation and showed that the candidate had some genuine curiosity in finding something worth talking about. Given that a candidate can code, and is comfortable with linear algebra, calculus and probability, they have the basics to learn everything else. It is very hard to teach someone to think creatively or become passionate about problems.
What is different about how Kissmetrics and Quick Sprout hire data scientists?
Right now Quick sprout is a very small team in the early stages of product development so we’re not currently hiring new data scientists. One thing that aspiring data scientists should know is that many startups and small companies are looking for a data scientist but may have given up on finding one as the search process can be exhausting. One of our best candidates at Kissmetrics showed up at our door and saying “I want to work here!”. People coming from academia or other large organizations might not be aware of how flexible startups and small companies can be when it comes to hiring. If there is a company out there you think is doing cool work, connect with them. It’s hard to make a better impression on a group of people excited about their work than telling them you love what they’re doing and want to be a part of it. Even if that company isn’t currently hiring you’ll be at the top of the list if when they do start looking.