(Editor’s note: this is part one of a two-part series. Part two can be viewed here).

Data scientists are known for their unique skill sets. While thousands of compelling articles have been written about what a data scientist does, most of these articles fall short in examining what happens after you’ve hired a new data scientist to your team.

The onboarding process for your data scientist should be based on the skills and areas of improvement you’ve identified for the tasks you want them to complete. Here’s how we do it at Elicit.

We’ve all seen the data scientist Venn diagrams over the past few years, which includes three high-level types of skills: programming, statistics/modeling, and domain expertise. Some even feature the ever-elusive “unicorn” at the center.

While these diagrams provide us with a broad understanding of the skillset required for the role in general, they don’t have enough detail to differentiate data scientists and their roles inside a specific organization. This can lead to poor hires and poor onboarding experiences.

If the root of what a data scientist does and is capable of is not well understood, then both parties are in for a bad experience.

Near the end of 2016, Anand Ramanathan wrote a post that really stuck with me called The Data Science Delusion. In it, Ramanathan talks about how within each layer of the data science Venn diagram there are degrees of understanding and capability.

For example, Ramanathan breaks down the modeling aspect into four quadrants based on modeling difficulty and system complexity, explaining that not every data scientist has to be capable in all four quadrants—that different problems call for different solutions and different skillsets.

For example, if I want to understand customer churn, I probably don’t need a deep learning solution. Conversely, if I’m trying to recognize images, a logistic regression probably isn’t going to help me much.

In short, you want your data scientist to be skilled in the specific areas that role will be responsible for within the context of your business.

Ramanathan’s article also made me reflect on our data science team here at Elicit. Anytime we want to solve a problem internally or with a client we use our "Geek Nerd Suit" framework to help us organize our thoughts.

Basically, it states that for any organization to run at optimal speed, the technology (Geek), analytics (Nerd), and business (Suit) functions must be collaborating and making decisions in lockstep. Upon closer inspection, the data science Venn diagram is actually comprised of Geek (programming), Nerd (statistics/modeling), and Suit (domain expertise) skills.

But those themes are too broad; they still lack the detail needed to differentiate the roles of a data scientist. And we’d heard this from our team internally: in a recent employee survey, the issue of career advancement, and more importantly, skills differentiation, cropped up from our data science team.

As a leadership team, we always knew the strengths and weaknesses of our team members, but for their own sense of career progression they were asking us to be more specific and transparent about them. This pushed us to go through the exercise of taking a closer look at our own evaluation techniques, and resulted in a list of specific competencies within the Geek, Nerd, and Suit themes. We now use these competencies both to assess new hires and to help them develop in their careers once they’ve joined us.

For example, under the Suit responsibilities we define a variety of competencies that, amongst other things, include adaptability, business acumen, and communication. Each competency then has explicit sets of criteria associated with them that illustrate a different level of mastery within that competency.

We’ve established four levels of differentiation: “entry level,” “intermediate,” “advanced” and “senior.” To illustrate, here’s the distinction between “entry level” and “intermediate” for the Suit: Adaptability competency:

Entry Level:

  • Analyzes both success and failures for clues to improvement.
  • Maintains composure during client meetings, remaining cool under pressure and not becoming defensive, even when under criticism.

Intermediate:

  • Experiments and perseveres to find solutions.
  • Reads situations quickly.
  • Swiftly learns new concepts, skills, and abilities when facing new problems.

And there are other specific criteria for the “advanced” and “senior” levels as well.

This led us to four unique data science titles—Data Scientist I, II, and III, as well as Senior Data Scientist, with the latter title still being explored for further differentiation.

The Geek Nerd Suit framework, and the definitions of the competencies within them, gives us clear, explicit criteria for assessing a new hire’s skillset in the three critical dimensions that are required for a data scientist to be successful.

In Part 2, I’ll discuss what we specifically do within the Geek Nerd Suit framework to onboard a new hire once they’ve joined us—how we begin to groom the elusive unicorn.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access