Top Five Qualities of a Data Scientist

If you are a developer or founder of a startup, it can be difficult to know what to look for in a data scientist, or hard to tell if you even need one. In this article, we will explore the most important technical and nontechnical skills a data scientist should have in their back pocket, so you know what to look for in a potential hire. 

1. Technical Skills. If you are running a startup, it's essential to have people who can swap hats easily. A data scientist may come from statistics, physics, chemistry, biology, or beyond. For your first hire, I recommend someone with a few years of software engineering under their belt. This will save your development team time headaches. What should they know? The most popular data science tools found by a Crowd Flower report are: SQL, Hadoop, Python, Java, R, Hive, Mapreduce, NoSQL, Pig, and SAS. Your data scientist in question doesn’t need to know about all of these, but it should be a huge red flag if they seem surprised when you bring one of them up in an interview. 
2. “Full Stack” Fluency. When you are first building out a team, you should look for the rare “full stack” data scientist or pair together a couple of “partial stack” data scientists. The full stack of a data scientist involves defining a business problem, collecting data, exploring data and analyzing it, creating Machine Learning models, deploying a model to make sure it is accessible, and monitoring the model. They should have experience in most of these, and can learn along the way if they have one or two weaknesses. 
3. Data Wrangling Prowess. The New York Times reported in 2014 that between 50%-80% of data scientists’ time is spent on collecting and cleaning data. It is also the most painful part of their job. In an interview, ask something along the lines of, “Talk about a time you collected and cleaned data, what were the challenges you faced and how did you push through them?” The more experiences they can point to, the more they’ve learned, the less likely they are to make costly mistakes. 
4. Soft Skills. I will say little about this, because it goes without saying that you need a team player if you want your team to be successful. I would prioritize soft skills over seniority all else equal if you are hiring one data scientist. This is because senior data scientists who can’t communicate well are often too specialized and better fit developed data science teams. Having someone who can communicate effectively likely means they will be able to work better with your development team, which is your top priority. 
5. Unique Perspective. The point of having diversity is that people can bring different backgrounds which complement each other and lead to new ideas and opportunities. If you already have a data scientist on your team, look for someone who has a different academic background. Don’t “hire for culture fit,” because this can reinforce your confirmation biases.  

Hopefully these traits give you some ideas on where to start the hiring process. Another question you might ask is, do I even need a data scientist? You may not if you have a team of developers who want to learn. It is worth it to have a developer tinker with Telepath’s free tier AutoML platform. They might find that they can solve most problems themselves, or realize that you need a data scientist’s help, which ensures they are worth the investment. If you are interested in learning more about what Telepath has to offer, click here.