6 tips for hiring a data scientist

You might think that all data scientists have the same skillset, but as a rapidly emerging job I find that it covers many roles and functions and few single candidates have them all.

Here are my top tips to help you qualify whether or not a data scientist is the right one for you.

1. Identify which type of data scientist they are

Here are the most common types of data scientists that I have encountered:

  • Data Engineer: a skilled programmer has the statistical knowledge and analytical, scientific mind set but typically lacks business knowledge and communication skills
  • Business Analyst: a business analyst has business awareness and a broad knowledge set, but lacks programing skills
  • Storyteller: a skilled visual communicator knows how to weave data and analysis into an easily understandable narrative; but lacks the programming, statistical and business experience to do the initial analysis

image01

With data science still in its infancy, it’s hard to find one data scientist that has all the required experience and capabilities. Finding someone who can meet your most pressing challenges now but then could grow into a wider role could be a more effective approach.

2. Consider developing talent internally

Recruiting a star data scientist is a lengthy, difficult and expensive process, so sometimes it can make more sense to develop internal talent. Whether it’s an experienced data engineer, programmer, or recent graduate you could already employ someone who could quickly develop into a data scientist with the proper training.

If you decide to develop talent internally, it might be best to start from statistical knowledge, and then add programming; as creativity will develop with confidence, then business knowledge and communication skills can be learned to become a well-rounded data scientist.

3. Build a team

Not only are data scientists with the full suite of capabilities expensive, they are also rare. The recruitment process could be very long, and there is always the risk that a data scientist will move on to a new challenge after being on board for a short while.

Rather than looking for a single individual that has all the needed skill sets it’s possible to build a team that performs the same function. A data science team that is an amalgamation of several people each with their own specialties can be a preferable solution.

Source: Hortonworks

Source: Hortonworks

Qualify you’re right for them

When you’re under pressure to hire a data scientist and you have found a strong candidate, you’ll want to give them an offer immediately before they go elsewhere.

Before you do, take the time to qualify that their data science experience fits your profile and that they understand the data challenges facing companies of your size in your industry.

Humans or computers?

There are two main types of data science “customers”: computers and humans. Depending on whether your data scientist will be producing analysis for machine learning algorithms, for example, or in order for a human to read and make a decision, they will need very different skills.

For example, a creative business analyst who understands how to construct a compelling narrative would be better at producing data for humans that a statistician who has poorer communication skills.

Standalone or embedded?

Another important consideration is whether your data scientist will be working in a separate data science team, or embedded within a different department.

Both of these approaches have their advantages: standalone teams offer more potential for collaboration and upward mobility; whereas embedded teams tend to create better alignment with project requirements. It’s important to ensure your prospective data scientist will be comfortable in whichever system you use.

Look for concrete accomplishments

“Data scientist” has been identified as the most exciting emerging job, so it’s understandable that everyone wants to become one. By looking for the following concrete examples of how your applicants have previously improved business processes, you will be able to identify if they have the necessary skills:

  • Using data to improve business processes and outcomes shows business knowledge

  • Following a scientific, rigorous process shows the right mind set

  • Ability to articulate past experience in an easily understandable way to construct a narrative around business process improvement.

Show they will be appreciated

Data scientists are one of the most in-demand skill sets on the market, and that means you will need to sell your company to them just as much as check whether they are right for you. This doesn’t just mean high pay: the best data scientists typically look for the most challenging projects and environments where they have the freedom to work effectively.

Therefore, being able to offer a wide range of projects and no micro-management will help you hire and keep the best data scientists.

Dave Akka, CEO ABRS, smart professional services and recruitment search