A research team,
not a generic AI.
A generic AI gives a plausible answer that can never be verified. Behind every Skillberg diagnosis, there is a research and engineering team and our own skills database, built since 2022.
The best studies, applied to your own tasks.
We match each task in your job descriptions to its equivalent in O*NET, the US occupational database used by Stanford and Anthropic. Each match gets a confidence level that is calculated, not declared.
So for each job, you know what the studies say and how many tasks each figure is based on. See the method in detail.
A skills database we built ourselves.
reference skills
Skillberg database
jobs covered
Skillberg database
place among organizations at TalentCLEF 2026, an international competition on matching CVs with job offers
TalentCLEF 2026
One skill opens the way to others.
Traditional frameworks are lists. They say nothing about what you need to know beforehand, or about where a skill leads. It's the difference between a list of cities and a road map.
Our skill trees, created in 2022, link each skill to the ones you need to master before it. They organize our database, which brings together five major frameworks, including ESCO (European), ROME (France Travail) and O*NET (US), and three others being integrated (Spain, Italy, Portugal).
Universities and public bodies by our side.
Université catholique de Louvain
The algorithms that link skills to one another.
Université catholique de Lille
Research and teaching partnership.
ARACT Hauts-de-France and Institut Godin
Two years of research on the impact of AI at work, with Z - Code pour l'emploi.
Thirty minutes to ask the right questions.
You tell us about your situation, we show you the tool in action on a fictional company, and together we sketch what a pilot would produce for you. No commitment.
30 minutes on video with Skillberg, no commitment. No personal data required.
- “About us” page (milestones and partners)
- Martin Vielvoye's post on the launch of the ARACT project, February 1, 2025