Research Engineer – Human Influence

Highlights

• Build RL environments to mitigate deceptive behavior in AI conversations.
• Run rigorous human-AI interaction studies and trials on influence risks.
• Build scalable evaluations tracking persuasion capabilities across releases.

Ideal candidate

A versatile researcher-engineer who loves building real systems and cares deeply about AI safety. You think across disciplines and communicate clearly about how AI shapes human beliefs.

Requirements

• Proven experience deploying a benchmark, evaluation, or product to users
• Clear understanding of current AI safety literature with interest in Human Influence
• Clear and consistent communication
• Clear understanding of fundamental Machine Learning concepts
• Experience fine-tuning or post-training LLMs using standard methods
• Comfortable with RL environments and reward-based methods for post-training models
• Able to write scalable, maintainable production code in Python
• Comfortable serving, scaling and containerising ML code on large compute clusters
• Must undergo criminal record check and obtain BPSS clearance; CTC eligibility preferred