Your search for Data Science Jobs

There's no single path in. Research-heavy roles still favour a doctorate in a quantitative field, and in the Netherlands PhD candidates are typically employed by their university, which makes the route into industry research roles well established. Away from that end of the market, a university (WO) or applied sciences (HBO) degree in a numerate subject, such as econometrics, mathematics, statistics, computer science, artificial intelligence or physics, covers most roles, and a working portfolio increasingly matters more than the qualification itself. Dutch universities of applied sciences also offer dual (duaal) routes that combine study with paid work inside a company, and if your qualification was gained abroad, it's worth checking how it's recognised in the Netherlands.

Whichever your route, the baseline is SQL, Python and genuine statistical grounding, built in that order. A portfolio matters more than most people expect and works alongside a certification rather than instead of one: a certification helps a CV through the earlier, often HR-led screening stage, while two or three real projects, built on messy, real-world data you can talk through in detail, carry the weight at interview.

Statistics and a genuine grasp of machine learning theory are what separate a data scientist from an analyst who's picked up some Python. What gets a CV shortlisted at mid-to-senior level is whether that theory turns into a working program that survives contact with real data. These are the skills that come up repeatedly across the data science roles we fill in the Netherlands:

• Statistics and probability - the mathematical foundation everything else sits on; still the single biggest differentiator at interview.
• Python - the standard language for building, testing and deploying models.
• Machine learning frameworks - scikit-learn, PyTorch and TensorFlow remain the most requested, with newer LLM-focused frameworks appearing in more job specs each year.
• Data analysis and experimentation - exploratory analysis, A/B testing and the judgement to know when a result is real rather than noise.
• SQL and data manipulation - pulling and shaping the data a model needs, rather than relying entirely on someone else to hand it over clean.
• Model deployment and MLOps - packaging a model so it runs reliably in production, not just in a notebook.
• Cloud platforms - AWS, Azure or Google Cloud experience is expected at mid-to-senior level, particularly for anything touching production machine learning.
• Data quality and validation - spotting a broken or biased dataset before it quietly wrecks a model's output.
• Privacy and governance - working within the AVG alongside the EU AI Act's obligations for higher-risk systems, most visibly in financial services, insurance and the public sector.
• Communication - translating a model's limitations honestly to the people making decisions based on it, and in client-facing and financial services roles, sometimes in Dutch as well as English.

Learning doesn't stop once you're in the role. The field moves fast enough that a data scientist who stopped updating their toolkit two years ago is already behind, and the candidates we place fastest are the ones who can point to something they've taught themselves recently, not just what a previous employer trained them on.

Dutch data science pay builds steadily with seniority, from the high €30,000s at entry level to about €50,000 for an established scientist, around €70,000 or more at senior level and over €100,000 in leadership. Machine learning and deep learning depth lifts the senior range, and freelancers typically quote between €94 and €120 per hour depending on the specialism. Here’s a breakdown guide per level:

Amsterdam has the largest concentration of data science employers and the highest advertised ceiling, while Rotterdam, Utrecht and Eindhoven add logistics, energy and high-tech work. With hybrid working standard, choose the specialism and sector you want first and the city second.

Data science careers fork later and less cleanly than data engineering, partly because the underlying skillset - statistics, programming and communication - transfers into so many adjacent roles.

• Junior to mid-level - moving from well-defined problems with a lot of oversight to owning a modelling project end to end, including the unglamorous parts like data cleaning and stakeholder buy-in, typically over two to three years.
• Mid to senior - the point where technical skill alone stops being the differentiator. You need to pick the right problem to solve, not just solve the one you're handed, and defend a modelling choice to people who won't read the maths.
• Senior to lead or principal - setting technical direction across multiple projects, often without formal management responsibility, and mentoring less experienced scientists through their first production deployment.
• Technical track - Principal Data Scientist or Applied Scientist, for those who'd rather keep building models than manage a team.
• Management track - Data Science Manager or Head of Data Science, for those drawn to shaping strategy and running a team rather than writing code day to day.
• Freelance: experienced scientists often move into ZZP or consulting work, an established route in the Dutch data and AI market that needs careful structuring under the Wet DBA.

Professional development carries more weight in this specialism than in most others we recruit for. Mid-level data scientists, find it's one of the benefits they weigh most heavily when comparing offers - unsurprising, given how quickly the underlying methods move.