Which qualification do you need for an AI job? What 400 Belgian vacancies really ask for
ikzoek editorial team · 4 min read · 4 October 2026
The number of computer science students in Belgium remains low, while demand for AI profiles keeps growing. Does that mean only computer scientists qualify for an AI job? We analysed the text of more than 400 active AI vacancies on ai-job.be and looked at which qualifications employers mention, and how often they mention none at all. The outcome is more nuanced than the cliché suggests.
What the vacancies ask for
| Mention in the vacancy text | Share of vacancies |
|---|---|
| A master's degree (general) | 54% |
| Computer science | 39% |
| Mathematics, statistics, physics or econometrics | 29% |
| A bachelor's degree is enough | 19% |
| A specific data science qualification | 10% |
| A doctorate | 6% |
| No qualification mentioned at all | 34% |
| "Or equivalent through experience" | 14% |
Computer science is therefore the most frequently mentioned field, but in almost all cases (153 of the 161 vacancies) it appears as one of several accepted qualifications, alongside engineering, mathematics, statistics or economics. Only three vacancies require exclusively a computer science degree. And a third of employers do not mention any qualification at all: they describe skills and experience instead.
The difference per type of AI job
The AI labour market is not a single whole. Roughly speaking, there are four families, each with a different profile:
- AI engineering and data engineering (together 62% of vacancies): building, integrating and maintaining models and data flows in production. Here computer science is mentioned most often (35 to 62%), because this is development work: Python appears in 41% of all vacancies, cloud platforms (AWS, Azure, GCP) in 36%, and MLOps tooling such as Docker and Kubernetes in 18%. An industrial or civil engineering degree is almost always listed alongside it as an alternative.
- Data science and machine learning (around 36%): analysing, modelling, experimenting. Here the picture is reversed: 41 vacancies mention mathematics or statistics without mentioning computer science. A quantitative background is more the norm here than a computer science degree.
- Applied and business roles: 38% of all vacancies feature a profile such as business analyst, functional analyst, product owner or consultant. They translate between the technology and the organisation and do not need to train models themselves.
- Specialised niches (NLP, computer vision, AI policy and ethics): small in number, with a doctorate or a specific master's degree as a more frequent requirement.
What employers are really looking for
Read past the qualification requirements, and three things stand out:
- Being able to programme, especially in Python, and being comfortable with SQL and version control. You can learn this in a computer science programme, but equally in an engineering or physics programme, a bootcamp, or simply by doing it for years.
- Statistical and mathematical insight: understanding what a model does and when it fails. This is the core of a mathematics, statistics or econometrics programme, and often precisely the weak point for many computer scientists.
- Domain knowledge and communication: knowing what the model is for, and being able to explain it to those who use it. That comes from experience in a sector, not from a qualification.
Routes to an AI job, by starting point
- You are still studying: computer science, civil or industrial engineering, mathematics, physics or business engineering are all good foundations. Choose elective courses in machine learning and data within your programme, and do an internship where you work with real data. A master's degree in AI (offered by KU Leuven, Ghent University, VUB, ULB and others) is a strong specialisation after a first master's degree.
- You already work in IT: the step towards data engineering or MLOps is small. Learn the cloud tooling and the basics of machine learning through online courses and certificates (AWS, Azure, Google), and build something you can show for it.
- You come from finance, marketing, healthcare or industry: your domain knowledge is scarce in AI teams. A postgraduate qualification in data science (one year, part-time possible) or a targeted course plus Python skills makes you a candidate for analytical and business roles. The skills radar shows, for each vacancy, which skills you already have and which ones you are missing.
- You do not have a higher education qualification: harder, but not impossible for engineering roles. A bootcamp, a GitHub portfolio with real projects, and a first job as a data analyst or junior developer are the usual path. A third of employers do not look at the qualification at all.
What this means for the shortage
The low number of computer science students is therefore not a wall for the AI sector, but it shifts recruitment: towards engineers, mathematicians and economists, towards career changers with domain knowledge, and towards people who prove themselves through experience. Employers who keep searching for a "master's in computer science with five years of AI experience" will search for a long time; those who open up the profile to skills find candidates faster. Read also skills instead of qualifications.
On ai-job.be you'll find all AI vacancies in Belgium, filtered by category (AI engineering, data science, data engineering, machine learning, MLOps, NLP, computer vision). Check the requirements per category, and compare them with what you can already do. There is a good chance you are closer to an AI job than the qualification cliché suggests.