If your organization is new to data science, chances are you are dealing with more unknown unknowns than known knowns as you build your data science team.
Here are some pointers to help you make the most of data science by hiring great data science professionals.
Hire Data Scientists for Passion
At a basic level, hiring data scientists is the same as filling any other role: You want people who are excited to come in every day and do great work. As the saying goes, you can teach skills, but you can’t teach passion.
Passion is critical for a data scientist because data is messy, results take time and there are lots of blind alleys to navigate before reaching the end of the rainbow. Getting there takes serious grit. Passion provides fuel to push through all the barriers, identify key data features and get results.
Good ways to check for passion include finding out how people became interested in data science, what projects they have done in their spare time and what they read.
People offer a broad range of answers when asked how they became interested in data science, and those answers are often entertaining. One person may have wanted to learn about image recognition, another may have wanted to master the math involved in playing poker.
Responses to the spare time question can also offer interesting insights.
Because data science is such a broad discipline, this question also elicits a wide range of positive answers. Many people may say that they take part in competitions on the Kaggle website, an online community for data scientists and machine learning engineers. Don’t be dissuaded if candidates talk about app and software development rather than data. Such activities still show the passion and grit you are looking for.
Asking people what they read is a great way to gain perspective on their level of technical understanding and their ability to communicate. For example, someone who reads FiveThirtyEight.com regularly comes across great examples of how to create compelling presentations of technical analysis for lay people. Toward the technical end of the spectrum are sites like the blog written by Columbia University professor Andrew Gelman, which dives into many questions of Bayesian statistics while still highlighting many of the challenges of communicating statistical data.
Ask About Failures
Real world data science is very different from classroom data science. In the classroom, students are handed pristine data sets chosen for their pedagogical power. Classroom data sets are designed or chosen to teach concepts and to help students gain an understanding of algorithms and techniques, not to show the daily challenges that a data scientist faces on the job.
Learning Opportunities
Webinar
Mar
21
Ready to Level Up Your Learning? Discover Your Learning Maturity Score
Learn how to level up your learning program in this free, interactive workshop
Webinar
Mar
29
The CX Commute vs. the CX Joy Ride
Is your CX traffic holding up your customer's journey?
Webinar
Mar
29
Is Your Microsoft 365 Data Recoverable? How to Survive the IT Hotseat When Data is Lost
See why MIcrosoft recommends third party backup
Webinar
Mar
30
An Optimized Customer Contact Strategy Combines Transparency and Intelligence: The State of Outbound Communications
Learn from the Neustar-commissioned Forrester Consulting study on challenges and solutions for contact centers
Conference
May
10
CMSWire CONNECT Customer Experience Conference - Austin 2023
Don't miss the top customer experience and digital experience conference of the year — live in Austin, Texas May 10-12.
Conference
May
10
Reworked CONNECT Employee Experience & Digital Workplace Conference Austin 2023
North America's best employee experience and digital workplace conference of the year — live in Austin May 10-12, 2023.
Common causes of failure, particularly for new data scientists, are starting with the wrong algorithm for the problem at hand, neglecting to clean and quality-check the data or not understanding the data, which can result in output giving false or misleading results.
Nobody likes to talk about their failures. My favorite way to work around this is to ask about key learnings they pulled from a project. For example, a clustering project I worked on ran the gamut from providing so few clusters that the resulting groups were obvious to providing so many that the model output was meaningless.
Don’t Be Overly Concerned About Tools
Tools change. Underlying concepts don’t. A carpenter understands the differences between a hand saw, a jig saw, a circular saw, a table saw, a radial arm saw, a reciprocating saw and a band saw. Fundamentally they are all just saws.
It is the same with data science tools. Whether the tool is SAS, R, the latest and greatest Python library or this week’s silver bullet of applications that will do all your data science for you — they are all just tools. They all use the same algorithms to evaluate data. Just the interface changes.
At the end of the day, data scientists need to understand data cleaning and algorithm choice and be cognizant of the project’s end goal.
Putting It All Together
You already have a great hiring process. Adding these tips to your arsenal will help ensure that you find the right people for your data science team. Happy hiring!
Learn how you can join our contributor community.