Bringing Order to VDAB's Growing Database

The challenge of a growing database

VDAB maintains a comprehensive database of over 11,000 competencies, skills, and knowledge elements across Flanders. This database is essential for connecting job seekers with suitable employment opportunities. However, as the database grew, manually grouping similar competencies and avoiding duplicates became an increasingly impossible task.

‍

‍

‍

An intelligent solution for matching competencies

The experts at Brainjar stepped in to develop a custom AI-powered solution. Their approach involved several steps:

1. First, they used the Google Translate API to create four language versions of each competency (Dutch, French, and English).

2. Then, they used advanced natural language processing techniques like word embeddings and sentence embeddings to analyse the meaning and relationships between different competencies.

3. Finally, they ranked the similarity of these competencies and implemented a feedback loop that allows VDAB to continuously refine the algorithms and ensure accuracy.

‍

‍

‍

A user-friendly interface for ongoing improvement

To make the solution as user-friendly as possible, Brainjar created a custom interface that allows VDAB's team to easily provide feedback and update the algorithms. This ensures that the competency matching remains accurate and relevant over time.

Thanks to this innovative solution, VDAB can now quickly and easily identify similar competencies within their database, saving countless hours of manual work and improving the overall efficiency of their operations.