Job description
White Circle • Paris or London
Paris, France, London, UK
$150,000 - $250,000
- In this role, you'll build and maintain an evaluation benchmark suite for AI safety, covering single/multi-turn content and agentic guardrails.
- Develop benchmarks that distinguish specific model capabilities and detect realistic failure modes.
- Collaborate with the product team to create evals for flagship models and emerging research features.
- Adapt evaluation frameworks across new verticals and evolving product requirements.
- Research and quantify agentic and language model failures in production environments.
White Circle helps companies test, protect, observe, and improve AI interactions.