Internships and graduation in AI: The complete guide for HBO and WO students

The transition from the lecture hall to the workplace is one of the most important steps in your career. Within the field of Artificial Intelligence (AI) and Machine Learning (ML), this step is perhaps even more crucial than in other fields. The theory you learn during your studies is valuable, but you only truly learn to understand the reality of messy datasets, legacy systems, and business interests on the work floor.

In this guide, we discuss everything you need to know about doing an internship and graduating in AI in the Netherlands. From finding the right workplace and formulating a feasible graduation project, to employer expectations and ways to turn an internship into your first real job as an AI professional.

Work-along internship versus graduation project in Artificial Intelligence

Depending on your academic year and study program, you will be looking for either a work-along internship or a graduation internship. Although both take place within a company, their goals and structures are completely different.

The work-along internship: gaining practical experience

During a work-along internship (often in the third year of an HBO program), you participate as a fully-fledged junior team member in an existing (data) team. You will be assigned tasks that directly contribute to the company's ongoing projects. Think of cleaning data, building simple data pipelines, training existing models with new data, or assisting with the implementation of an API for a Large Language Model.

The goal here is to gain work experience, get to know business processes, and experience what it is like to work in an Agile/Scrum environment. You are expected to be eager to learn, ask questions, and write code that is reviewed by senior developers.

The graduation internship: solving a problem independently

A graduation internship revolves around a specific research question or a defined project that you carry out independently from start to finish. Instead of participating in daily operations, you act almost as an internal consultant. You identify a problem (or the company already has one ready for you), conduct literature research, build a prototype or model, and validate whether this model actually solves the problem.

For a graduation project, it is essential that the scope is clear and that the project does not depend on processes over which you have no influence (such as data that still needs to be collected).

The difference between HBO and WO internships in AI

Both HBO and WO students are highly sought after in the AI job market, but employer expectations often differ based on the nature of your educational background.

The HBO student: Applied AI and Engineering

HBO students (for example, from programs like Applied Mathematics, HBO-ICT with a specialization in AI, or Data Science) often excel in practical application. Companies expect an HBO intern to become operational quickly. The focus is less on mathematically optimizing a neural network from scratch, and more on successfully implementing existing machine learning techniques and open-source libraries to solve a direct business problem. You are often a bridge between theory and the software engineering side.

The WO student: Fundamental research and complex models

WO students (from programs such as Artificial Intelligence, Computer Science, or Data Science at universities) are more often deployed on complex, fundamental issues. Think of modifying a model's architecture, in-depth statistical research into data bias, or solving problems for which no standard 'out-of-the-box' solution yet exists. Here, mathematical foundation and methodologically correct research carry more weight.

Where do you find the best AI internship positions in the Netherlands?

The Netherlands has a thriving tech ecosystem. Where you will fit in best depends on your ambitions and working style.

Tech hubs and ecosystems

Large concentrations of AI companies can be found in specific regions. The Amsterdam Science Park is a breeding ground for AI, partly due to the presence of major institutes and the UvA/VU. In the Eindhoven region (Brainport), the focus is heavily on high-tech, robotics, and embedded AI (for example, at ASML or Philips). In cities like Delft and Enschede, you will find many spin-offs from the technical universities, often focused on deep-tech and medical AI.

Startups, scale-ups, or corporates?

• Startups: Ideal if you want a lot of freedom, have broad interests, and don't mind that processes are still chaotic. Your impact is directly visible, but guidance can sometimes be scarce.
• Corporates: Banks, insurers, and large retail companies often have massive amounts of data. Here you will learn how to deploy AI at scale, how data governance works, and you will often receive professional, structured guidance.

The (semi-)government and non-profit

An often overlooked sector is the government. Municipalities, ministries, the police, and agencies such as the CBS (Statistics Netherlands) and the Tax and Customs Administration have a huge need for AI talent. Internships here often revolve around societal impact, fraud detection, predictive models for urban planning, or ethical AI. Read more about the unique opportunities in our article on AI roles in government.

What do companies expect from an AI intern?

Companies look for a balance between technical skills and communication skills. Rest assured: no one expects you to be a senior expert, but the basics must be in order.

Technical foundations (Hard skills)

Make sure you are fluent in Python. This is the de facto standard language. Knowledge of SQL is also an absolute must; you will need to be able to extract your own data from databases. Furthermore, you are expected to be familiar with the standard stack: Pandas, NumPy, Scikit-learn, and ideally a deep learning framework like PyTorch or TensorFlow. For LLM-related assignments, basic knowledge of frameworks like LangChain or LlamaIndex is a big plus.

Communication and domain knowledge (Soft skills)

The biggest frustration for managers is an intern who builds a brilliant model but cannot explain its business value or how it works. Models must be interpretable by 'the business'. Stakeholder management, presenting, and translating jargon into understandable language are essential. Discover how to train these skills in our guide on soft skills in AI.

How do you formulate a strong AI graduation project?

A poorly formulated project description inevitably leads to study delay. Follow these steps to write a watertight graduation proposal:

Step 1: Determine the business problem, not the technology

Do not start with: "I want to build a neural network." Start with: "The company loses 10% of its customers monthly and we do not know why." The AI technology is merely the tool to solve the problem (predicting churn). If you approach the project from a technology perspective, you risk building a solution for a problem that does not exist.

Step 2: The Data Check (The ultimate pitfall)

Before you sign off on a project, you must demand to see the data. Is the data labeled? Is there enough historical data? Can the data be used under the GDPR? If the answer is: "We will collect the data during your internship," you should decline or reformulate the project. You do not have time to wait three months for data.

Step 3: Define success and baselines

How do you know at the end of your internship if your project was successful? Ensure there is a measurable baseline. What is the current error margin of the manual process? Which metric will you use? (e.g., F1-score, RMSE, or a financial saving). Agree clearly on when the model is 'good enough'.

From internship to permanent job: how to make yourself indispensable

An internship is essentially a five- or six-month job interview. Many companies explicitly use internships to attract young talent. How do you ensure you can stay after graduation?

Build a tangible portfolio

Make sure the code you write is modular, documented (think of docstrings, typings, and a good README), and transferable. If you leave after your internship and your code dies a quiet death on a forgotten server, you have made little impact. By writing robust software, you demonstrate seniority. It is also wise to keep anonymous or open-source versions of your work during your internship. Read more here about building a strong AI portfolio.

Take initiative outside of your project

Do not focus exclusively on your graduation thesis, however important it may be. Participate in company hackathons, join internal data guilds or knowledge sharing sessions, and show interest in your colleagues' work. Visibility is crucial.

Prepare for the transition

If you are in the final month of your internship and a contract has not yet been discussed, proactively bring this up with your manager. Show that you know your market value and what role you would like to fill. Do you want a seamless transition? Make sure you put the tips from our article on applying for an AI position into practice during your internship.

Frequently Asked Questions about AI Internships (FAQ)

Do you always receive an internship allowance for an AI internship?

Yes, in almost all cases. Because AI talent is scarce, many commercial companies pay above-average internship allowances. In the Netherlands, this usually varies between €400 and €800 gross per month. At (semi-)governmental organizations, these allowances are often laid down in a collective labor agreement (CAO) and are around €400 to €550.

Do I already need to be an expert in Generative AI (LLMs)?

No. Although the term AI nowadays often seems synonymous with ChatGPT, most data projects at Dutch companies still consist of 'classical' machine learning (regression, classification, predictive maintenance). Knowledge of Large Language Models is a great addition, but solid knowledge of traditional machine learning and good data analysis is at least as important, if not more so.

Can I do an internship at a company without a data science team?

This is possible, but very risky, especially for a graduation internship. If you are the only one with knowledge of AI in the company, no one can guide you technically, review your code, or help you when you get stuck on complex mathematical or architectural problems. For your own learning curve, it is highly recommended to look for an internship at a company with at least one, but preferably several, experienced data scientists or AI engineers.