About Veridak
Building the foundations of French-language AI.
Why Veridak exists
AI models are trained almost exclusively on English data. Less than 5% of training content is in French, and virtually none comes from Quebec domain experts.
The result? Models that invent legal articles, hallucinate on Canadian medical standards, and confuse French and Quebec law. For companies deploying AI in critical contexts, that's a real risk.
Veridak was born from this observation. We connect French-speaking professional experts with AI labs that need reliable data to improve their French-language models. Lawyers, doctors, engineers, accountants: our experts create professional-quality data.
The market at a glance
Our mission
Create professional-quality French-language AI training data, produced by real experts, within an ethical and legally irreproachable framework.
Our vision
Become the reference for expert French-language data for AI labs worldwide. So that every AI model speaks precise, contextual and reliable French.
Our values
Rigor
Every data point goes through double quality control. Our experts are certified professionals who create original content.
Transparency
Complete traceability for every data point produced. You know exactly where what you're buying comes from.
Ethics
Fair expert compensation, regulatory compliance, zero scraping. We do things right.
Impact
We're improving AI for 300 million French speakers. This isn't just a business, it's a real need.
The team
Based in Quebec. Data science, engineering and project management.
Camelia Raymond
Co-founder, Data Science & AI
Senior Data Scientist, specialized in trustworthy AI and explainability (XAI). Master's in Trustworthy AI.
Arthur Villaret
Co-founder, Project Manager
Biomedical engineer with expertise in complex technical project management.