Introducing WorkRB: An Open Benchmark for AI in the Work Domain

Matthias De Lange
December 8, 2025
3 min read
Contents

AI models are transforming how we understand jobs and skills, but progress is slowed by scarce and scattered public data. Today, information is fragmented in isolated datasets and ontologies, making it hard to compare results or build on each other’s work. In a rapidly evolving AI landscape, this fragmentation holds back collective advancement. WorkRB aims to fix this by creating a shared research ecosystem where every contribution strengthens many others.

WorkRB, you say? 🐝

The Work Research Benchmark (WorkRB) is a new open-source evaluation toolbox that brings everything you need into one place. It bundles datasets, ontologies, and baseline models in a single, easy-to-use toolkit, so anyone can quickly assess how well their model understands jobs, skills, and work related tasks. Its modular design makes it simple to scale across many tasks and languages, giving researchers and practitioners a shared foundation.

What you can do with WorkRB

The toolkit covers a wide range of evaluations, including job-skill matching, normalization, extraction, and similarity. The built-in multilinguality aims to reflect the wide spectrum of languages in today’s global workforce. With ready-to-go baseline models and extensible interfaces, WorkRB lowers the barrier to entry for anyone who wants to evaluate their AI model capabilities.

Community-driven!

As the field evolves at high speed, building shared foundations is a continuous effort. We expect strong momentum in the coming months as contributors from all corners of the ecosystem enrich the benchmark with new datasets, ontologies, and models. To catalyze this growth, Techwolf is launching WorkRB alongside a community paper that welcomes impactful contributions from academia, industry, and government. Our goal is to unite these communities around a common benchmark that elevates the entire field of work related AI.

Join the effort

With WorkRB, we hope to build a trusted, open benchmark for work related AI that accelerates progress across the entire ecosystem. This is only the beginning, and we invite everyone interested in the future of work and AI to take part.

Want to contribute? Reach out to workrb@techwolf.ai 

Explore the project

Discover here
Discover here

Blog

Relevant sources

From guides to whitepapers, we’ve got everything you need to master job-to-skill profiles.

View all
View all
AI
Blogpost

Introducing WorkRB: An Open Benchmark for AI in the Work Domain

WorkRB is a new open-source evaluation toolbox designed to fix this by bundling datasets, ontologies, and models into one shared foundation. We invite the community to join this effort to accelerate and elevate the field of work-related AI.
Matthias De Lange
Dec 8, 2025
Introducing WorkRB: An Open Benchmark for AI in the Work Domain
AI
Work Intelligence
Blogpost

The vision for a skill-first world: Introducing the TechWolf skill intelligence index 

Complementing our high-level Work Intelligence Index, the new Skills Intelligence Index offers a granular view of how AI impacts individual skills.
Jeroen Van Hautte
Nov 20, 2025
The vision for a skill-first world: Introducing the TechWolf skill intelligence index 
AI
Work Intelligence
Blogpost

How it works: The TechWolf skills intelligence index unpacked

Go inside TechWolf’s Skills Intelligence Index to see how we transform over 2 billion job postings into a comprehensive map of skills and tasks.
Jeroen Van Hautte
Nov 20, 2025
How it works: The TechWolf skills intelligence index unpacked

Using AI while interviewing at Techwolf

At TechWolf, we see generative AI as part of the modern toolkit — and we expect candidates to treat it that way too. We love it when people use AI to take their thinking to the next level, rather than to replace it.You are welcome to use tools like ChatGPT, Claude, or others during our interview process, especially in take-home assignments or technical exercises. We encourage you to bring your full toolkit — and that includes AI — as long as it reflects your own thinking, decisions and creativity.We don’t see AI as replacing your skills. Instead, we’re interested in how you use it: to brainstorm ideas, speed up iteration, validate your thinking, or unlock new ways of approaching a challenge. Great candidates show judgment in when to rely on AI, how to adapt its output, and where to go beyond it.

What we’re looking for:

Our interviews are designed to understand how you think, solve problems, and express ideas. Using AI in a way that amplifies those things — not masks them — is encouraged.

What to avoid:

We ask that you don’t submit AI-generated work without review, or present answers that you can’t fully explain. We’re not testing the model — we’re getting to know you, your skills, and your potential. If there are cases where we don’t want you to use AI for something, we’ll tell you ahead of the interview being booked.In short: use AI as you would on the job — as a smart assistant, not a stand-in.

Example: Programming with AI

In a coding challenge, you’re welcome to use generative AI to support your workflow — just like you might in a real development environment. For instance, you might use AI to quickly generate boilerplate code, look up syntax, or get a first-pass solution that you then adapt and debug collaboratively. What we’re interested in is your ability to reason through trade-offs, communicate clearly, think about complexity and iterate effectively — not whether you memorized the syntax perfectly. If using AI helps you stay in flow and focus on higher-level problem-solving, we consider that a strength. There could be some challenges where we won’t allow you to use AI - in that case we’ll tell you in advance, and will tell you why.

Heading

Contents

Heading 1

Heading 2

Heading 3

Heading 4

Heading 5
Heading 6

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.

Block quote

Ordered list

  1. Item 1
  2. Item 2
  3. Item 3

Unordered list

  • Item A
  • Item B
  • Item C

Text link

Bold text

Emphasis

Superscript

Subscript