Research Scientist/Engineer (Fundamental Research)
Core
Conduct fundamental research on AI alignment, reward-seeking, and scheming risks to improve embedded evaluation methods for frontier AI systems.
Role type
Senior Research Scientist/Engineer (Fundamental Research)
Builds
Novel evaluation techniques, empirical foundations for scaling laws of scheming, and insights into AI cognition.
Domain
Artificial Intelligence / AI Safety / Alignment
Deliverable
research
Required skills
Reinforcement learning training of LLMs, experimental design and execution, quantitative analysis (scaling laws, statistical physics), conceptual thinking and scientific writing, AI power-user (prompt engineering, agent workflows), Python programming
Preferred skills
Project management, software engineering skills
Technologies
Python, LLMs, Reinforcement Learning
Responsibilities
Collaborate with leading AI developers across multiple labs; Study RL dynamics leading to reward-seeking and misaligned preferences; Build empirical foundations to predict how scheming risks evolve as models scale; Develop novel evaluation techniques for highly evaluation-aware models; Deep dive into AI cognition to discover new reasoning patterns.
