Hiring Data Scientists

Lessons from the trenches

Data scientist recruiting is a major pain point for many organizations, chiefly for three reasons:

  1. vague and hard-to-assess skill set
  2. lack of data scientists on staff who are experienced with recruiting and candidate assessment
  3. no clear “ivy league” backgrounds that it is safe to hire from.

Companies who master the art of data scientist recruiting can turn this into a significant competitive advantage: getting hands on top data science talent helps them to sharpen their competitive edge as data scientists are the key resource for any organization willing to spur growth, gain traction and succeed in a digitalized world. For service providers, such as management consultancies or digital agencies, excellence in data science recruiting may be a matter of survival – or a unique capability that commands a premium.

In a profession as new as data science, naturally, recruiting best practices are yet to emerge. This oftentimes leads to pre-mature hires and disappointments, when the new recruits do not perform as expected on the job – a situation that is typically detected too late, and not acted on decisively. What’s worse, this might lead to a loss of trust into the data science function as such and the organization as a whole might challenge the value of in-house data scientists if no tangible results are delivered straight away.

Hiring a data scientist is one of the biggest challenges in recruiting, and will remain so until the size of the talent pool has increased significantly and the first cohort of data science leaders has grown up. In this white paper, we discuss five lessons learned that can provide some guidance for companies who want to calibrate their recruiting process for data scientists.


  • Assess your internal setup and clarify target roles
  • Clarify the envisioned data scientist skill set
  • Data Science domains and specificity of methodology
  • The absence of senior data scientists and how to deal with it
  • CVs are meaningless, there is no “ivy league” of data science
  • Show me your work: the right way to assess