The art of finding a good data scientist Posted on : Jul 04 - 2018

In an era when data processing and storage have become cheap and artificial intelligence is coming of age to deliver new insights, data scientists have become a white-hot commodity.

A great scientist understands both the science and math behind algorithms and is a deft computer programmer. This hard-to-find combination makes a top-notch data scientist rare in both startups and large organizations alike.

The demand for data scientists will increase 28 percent by 2020, according to a report by IBM. This surge is creating a shortage of workers with these specialized skills at a time when data-savvy organizations need them most.

To close the gap, companies should be imaginative about where they search for data scientists.

Two types of data scientist candidates

Data scientists tend to come in two flavors: The first group is people with traditional computer science backgrounds who later developed an interest in data science, and learned about overlapping technologies such as artificial intelligence and machine learning. The second group of people are scientists or data scientists first – men and women with hard-core training in mathematics, statistics, and algorithms, who have learned programming to support their research. 

In my experience, the best data scientists often come from the second group, yet most companies focus on the first during their candidate searches.

Now don’t get me wrong. Plenty of excellent data scientists started their careers as computer scientists with strong programming skills, and there are some great computer science programs at schools around the country. And companies, especially startups with only so many resources, must be mindful of the right balance between theoretically-oriented data experts and those adept at the programming needed to develop a product that a customer can use. Startups ideally want a person who understands the science and can productize the algorithms they create.

But by looking beyond the “usual suspect” talent pools toward straight scientists, organizations may attract thinkers with the creative curiosity and problem-solving orientation needed to achieve data science’s ultimate goal – finding out what is not known and turning those insights into strategic business action.

Also, it’s simply easier to take someone who is highly skilled at science and math and teach them how to program than the reverse, in my experience. I have two physicists and a mathematician on my team, and they’re strong contributors to our company’s artificial intelligence-driven mission. View More