Stuart Gentle Publisher at Onrec

How to interview a data scientist

A 2019 report by the Royal Society warned that demand for data scientists had more than tripled over five years, leaving businesses "crying out for professionals to unlock the potential of new technologies".

John Salt, the CEO of

If we were to look at the data science job market today, recent opportunities with the Royal Household, the England Rugby team and writing algorithms for a quantum computer tell us that demand hasn’t dwindled. But how many times have you, as a HR director, recruiter or personnel manager, interviewed a data scientist?

No matter your answer, I can guarantee that, irrespective of the industry or sector you recruit for, that number is going to increase exponentially. Across industry, companies are increasing the number of data scientists they hire to help manage, analyse and interpret the numerous datasets the continual digitalisation of industry and commerce creates.

In 2022, Statistica published a piece of market research, which surveyed 403 individuals with a level of insight into the machine learning efforts of their companies across a random sampling of industries and machine learning maturity levels.

The research showed that, in the twelve months of 2020 to 2021, the percentage of surveyed organisations that employed 50 data scientists or more increased from 30 per cent to almost 60 per cent. On average, the number of data scientists employed in an organisation grew from 28 to 50.

However, that doesn’t necessarily mean that the recruitment function is ready for this new wave of applicants. Data scientists do a unique job, using tools that are rarely used outside their sector and, crucially, the way they view the outputs of their job is highly technical and complex.

Our task in recruitment is to understand those outputs and evaluate their suitability for the role in hand. Fortunately, there are steps you can take to ensure you get meaningful information out of the process, which will be valuable for all levels of the business.

Look for stories with actionable data in them

Do you really care whether the candidate prefers Snowflake or Splunk? Do you even know what that means? A much more relevant story would be one involving actionable organisational improvement that came as the result of the candidate’s work.

How did they improve an algorithm or model and what did that mean for the business, for instance? Did the work they do result in a more sellable or more sold product or service? Did it lead to staff retention, or even, in the case of the England Rugby, more effective team play?

Remember, having the right credentials is one thing, proving that the data scientist can use those credentials to create real, lasting business value is the true story recruiting companies want to hear.

Look for the candidates that start with challenges, not data

The interviewee should be excited by data, but more importantly they should be excited by the problems they can solve using that data. Look for clear evidence of a solution orientated mindset. Ideally, the candidate will start with examples of the problems they have been called in to solve in past roles.

For instance, the job in the Royal Household was about producing and analysing project data on the wing-by-wing overhaul of the Palace’s infrastructure. So here, you would be looking for evidence that they understood and embraced the purpose of this project, not just that they are comfortable in Power BI and Azure.

Do a little bit of research

If you don’t know your MLOps from your deep learning or your Azure from your AWS, you should probably do some initial research before the interview, to ensure you understand the answers and the candidate’s requirements.

You don’t want to onboard your first data scientist and then discover they can’t do their job because they don’t have the right tools. Medium has a channel called Towards Data Science that would be a great starting point, especially this article by Ignacio Montegu on the foundations of the industry.

These three steps provide the perfect preparation for your next data science recruitment project, whether you need to recruit a quantum computer programmer, find the data scientist for a future season of The Crown, or help England land the Six Nations.

About the author: John Salt is the CEO of, the UK’s largest data science only jobs board. An industry veteran, with time at TotalJobs Group, C-V Library, Reed Elsevier and Guardian Media, he’s an expert in internet marketplace, e-commerce and SaaS business models. He built Totaljobs Group sales channels and e-commerce from less than £2M to more than £70M, delivering profitable growth both there and at CV-Library.