Technical Programme Manager - Model Transparency

London, UK

The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.

We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.

AISI aims to turn progress on important research questions in AI security into real insight and impact for our partners and the public. This requires running research programmes that actually deliver, and making these programmes rigorous and fast is its own discipline. The AI Security Institute is hiring a Technical Programme Manager for our Model Transparency team to do exactly this.

Over the next twelve months, this team will need to move faster, deliver more complex research programmes and evaluations, and engage more deeply with partners in major AI labs and across government than ever before.

Technical Programme Managers make this work possible. You will work directly with the team’s researchers and engineers to turn vast research questions into tractable programmes. You will be held responsible for ensuring our research stands up to rigorous empirical standards to inform policy, recommending which technical work needs to happen, and untangling knotty dependencies across work streams.

The ability to effectively evaluate and monitor AI systems will grow in importance as models become more capable, autonomous, and integrated into society. If models can detect and game evaluations, obscure their reasoning, or behave differently under observation, the safety claims that governments and developers rely on become unreliable. Understanding and addressing these risks is essential to ensuring that oversight of advanced AI systems keeps pace with their capabilities.

The Model Transparency team is a research team within AISI focused on ensuring that evaluations, assessments, and monitoring of frontier AI systems remain reliable as models become less transparent. We research how and why oversight is declining – through phenomena such as evaluation awareness, unfaithful chain-of-thought reasoning, and changes in model architectures – and develop methods (including white and black box methods) to detect, measure, and mitigate potential issues.

We share our findings with frontier AI companies (including Anthropic, OpenAI, Deep Mind), UK government officials, and allied governments, and publicly to inform their deployment, research, and policy decisions. We also work directly with safety teams at frontier labs, contributing to safety case reviews and helping improve their alignment evaluation methodology.

Our recent work includes the loss of oversight report, auditing games for sandbagging, reproducing natural emergent misalignment from reward hacking, and identifying open-weight language models that game propensity evaluations.

Programme architecture. You will work directly with your team's technical staff to set the scope of projects: refining what questions we're actually trying to answer, what the milestones and success criteria are for technical work, and how to sequence dependencies, resource, and time across the team. You will turn "we should probably look into X" into a tractable plan with clear deliverables, without imposing process that gets in the team's way.

You must be comfortable with ambiguity: you will be immersed in a…