Summary

Build evaluation datasets, reinforcement learning environments, and verification systems for frontier AI in finance, healthcare, and legal domains. This is a London-based role on a small, hands-on team working across engineering, evaluation design, and domain-expert collaboration.

Responsibilities

• Build reinforcement learning environments across regulated domains
• Design tasks with golden answers and calibrated rubrics
• Write verification logic and model reward functions
• Translate professional workflows into structured tasks
• Prototype evaluation and synthetic data approaches