login


Google Sign inLinkedIn Sign in

Don't have an account? 

Hiring your first data scientist without data science knowledge

Remote
Dev Processes
Hiring
Data team

26 November, 2018

Brad Pflum gives some advice on recruiting the first data scientist for a company, the procedures and the pitfalls.

Problem

With all the data that companies have right now, hiring a data scientist could be a future-proof asset in order to make more data-driven decisions. However, when recruiting a completely new data scientist in the staff, it can be quite tricky when no one is a specialist in this field. Your software engineers can certainly help in building a test to measure their capabilities in SQL and Python, but is a pure technical test enough to hire a data scientist?

Actions taken

The first questions to ask to yourself is: what do you want to get out of this person? What success will look like in the next 6 weeks, the next 6 months and the next 2 years? What stories and what conclusion do you want to be discovered out the data? Be ready to have a well formed job description because there are probably 60 different versions of data scientist, with some who code in R rather than in Python. The best first hire for a company in data science would be a full-stack data scientist, i.e. someone who works with the product owner or the commercial people to figure out the problem and the questions to be asked. The pitfall in this case is not to mix up the full-stack data scientist with a data engineer who takes care of creating databases, pulling and marshaling data, and bringing it into a data warehouse. Data scientists might struggle in these areas because data engineering requires a different skill set than data science. What might be needed for a company who wants to create a data section is to hire both a data engineer with analytics knowledge and a data scientist. If the tests are too focused on the technical skills, the data scientist candidates might drop all the way to operational databases and be frustrated not to be able to do anything. So after a short technical test, it would be best to assess the candidates on a cross functional interview. What you want to evaluate is their capabilities to find the right questions to ask, extracting insights out of the data and translate that into an accomplishment for the product/business counterpart. An example would be to take public dataset and see how they pull insights and show on this test their creative thinking and their problem-solving skills. If you have no one who can evaluate these skills, try to find an advisor/consultant who can test them on basic statistics and basic machine learning algorithms (classification, difference between supervised and unsupervised algorithms for example). Then the question that I always ask during the test is: how do you know your model is right? Usually they would analyze the residuals or inspect the errors in their machine learning model.

Lessons learned

To sum-up, here are the takeaways:

  • Pin down the questions you want your newly data science hire to answer
  • Have a well formed job description because there are different kinds of data scientists
  • Don't focus too much on the technical test: they might use R or Python or their skill set might not even be adapted for the test you make for them
  • You might want to hire a data engineer with analytics knowledge too
  • You want to evaluate them on a cross functional interview: extracting insights, creativity, problem-solving, self-evaluation
  • Hiring a remote data scientist might be tough, especially if you do not have a data department

Related stories

Recruitment and Interview Rotas: the Engineers’ Way
30 June

Jeff Foster, Head of Product Engineering, explains how engineers at his organization self-managed their taking part in the interviewing process.

Hiring
Collaboration
Jeff Foster

Jeff Foster

Head of Product Engineering at Redgate

Some Useful Tips for Decoupling Releases and Deployments
30 June

Pierre Bergamin, VP of Engineering at Assignar, outlines some useful tips for decoupling releases from deployment and increasing deployments by a huge factor, speeding up reverts and planning releases in a better way.

Agile / Scrum
Dev Processes
Pierre Bergamin

Pierre Bergamin

VP of Engineering at Assignar

Asynchronous Written Communication: How to Make It Work
30 June

Jose Pettoruti, Director of Engineering at CurrencyCloud, describes how to handle multiple inputs from a number of people in an asynchronous mode.

Remote
Internal Communication
Jose Pettoruti

Jose Pettoruti

Director of Engineering at CurrencyCloud

Deciding on a New Manager: What Really Matters
30 June

Jose Pettoruti, Director of Engineering at CurrencyCloud, explains what he considers when deciding on a new manager.

Hiring
Jose Pettoruti

Jose Pettoruti

Director of Engineering at CurrencyCloud

Why Psychological Safety Is More Important Than Ever During the Covid-19 Pandemic?
30 June

Murali Bala, Director, Software Engineering at Capital One, discusses the importance of psychological safety emphasizing its unparalleled significance during the Covid-19 pandemic.

Remote
Feedback
Team reaction
Feelings aside
Company Culture
Internal Communication
Murali Bala

Murali Bala

Director, Software Engineering at Capital One

You're a great engineer.
Become a great engineering leader.

Plato (platohq.com) is the world's biggest mentorship platform for engineering managers & product managers. We've curated a community of mentors who are the tech industry's best engineering & product leaders from companies like Facebook, Lyft, Slack, Airbnb, Gusto, and more.