Hire a data scientist

A data scientist investigates what the data says about a question your organisation has. They analyse data, build and test models, set up experiments and explain the results to the people making decisions.

On our secondment page, you’ll find data scientists under Data Science & AI, alongside roles such as AI engineer and machine learning specialist. Colleagues discuss the profession itself in De Databar, episode 10.

A data scientist asks: what relationships exist, what model can predict the outcome, and does it work better than the current approach? A machine learning engineer puts that model into production and keeps it running. A BI specialist creates reports and dashboards using data that is already available.

In small teams, one person often covers more than one of these roles. In our first conversation, we’ll work out which profile fits.

You contract a freelance data scientist yourself, often through a platform. That means you assess their quality and whether the assignment complies with the rules for self-employed workers. With secondment through Data Science Lab, the data scientist is our employee. We handle selection, onboarding and follow-up.

You’re not on your own, either: the data scientist can draw on colleagues, our internal lab and the Techdays.

Since 1 January 2025, the Dutch Tax and Customs Administration has once again been fully enforcing the rules against false self-employment. Whether someone is effectively working as an employee depends on how the arrangement works in practice, including factors such as supervision, integration into the organisation and whether they operate as an independent business.

A data scientist seconded by Data Science Lab has an employment contract with us. That differs from a freelancer you contract yourself. Have HR, procurement or an adviser review your own hiring policy.

Usually Python and SQL, in notebooks on your existing platform, such as Databricks or Microsoft Fabric. Their work also involves statistics, regression and machine learning, as well as experiments with a control group. MLflow records which data and settings produced each result.

Just as important is translating the findings: explaining a result clearly enough for a doctor, planner or marketer to make a decision based on it. You can read how we build models on our machine learning page.

By deciding in advance what needs to improve and how you’ll measure it. You test a model on data it has not seen, such as data from a later period, and compare it with the current approach or a simple rule.

If you’re testing an intervention, such as a retention campaign, a control group shows whether the intervention caused the difference. Without that test, it remains an assumption.

Analyses, notebooks, code and documentation stay in your environment and repository, not ours. Where appropriate, the data scientist works alongside an internal colleague who will take over the work.

If the collaboration goes well and you want to bring that expertise in-house permanently, we can discuss hiring the data scientist as a permanent employee.

If your data is not yet in order or accessible, a data scientist has little to work with. Start instead with a data engineer or a data foundation. If you mainly need reports, a BI consultant is a better fit.

If you want a completed application rather than extra capacity, have it built as a project through data science.