Technical Programme Manager - Model Transparency

The deadline for applying to this role is Sunday 1st November, end of day, anywhere on Earth.

About the role

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 workstreams. 

About the team

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, DeepMind), 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.

What you will own

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 fast-moving environment and won’t always have a settled roadmap. . 

Delivery tracking and unblocking. You will spot and solve problems so that the team's delivery continually improves, and own tracking progress to create visibility and establish control without overhead. You will surface technical blockers before they become crises, intervene directly to resolve them, and know exactly when and to whom to escalate. The goal is the team spending their time on the work, not logistics. 

Cross-government technical engagement. You will speak with government partners on a technical level and be a trusted representative and ambassador for good relationships with them.  

Team operations and pace. You will determine the operational rhythm that lets a small, high-performing team inside government move at the speed of billion-dollar AI companies. You will ensure the team has what it needs to deliver and be unfussy about doing whatever needs to be done to unblock things and prevent them from falling through the cracks. 

Role Requirements

• A track record working with technical colleagues. You have delivered technical programmes in a similarly high-ambiguity environment. You know how to scope and structure open-ended goals (not just deliver against a pre-defined plan). You have earned the trust of technical teams by understanding their work deeply enough to provide critical challenge and engage beyond process.  
• Comfort operating at the frontier of AI. You do not need to be a machine learning researcher, but you need to be fluent enough in frontier AI to operate in constant conversation with technical staff about our work, strategy, and research taste; you should have your own taste on what's hard, what's speculative, what's blocked, and what "good" looks like for a given experiment or evaluation.  
• High agency and ownership. You take responsibility for outcomes, not activities. When something is stuck, you pick up the phone. You do not wait for someone else to escalate.  
• Credibility with technical teams. Researchers trust you because you understand their work, respect their constraints, and demonstrably make their lives easier. You influence through competence and service, not authority or process.  
• Strong prioritisation under pressure. You can identify what actually matters on a daily, weekly, monthly, and quarterly level and focus the team's energy. You have the judgement to recognise work that isn't landing and the confidence to recommend that it stop.  
• Discretion and judgement in a sensitive domain. You understand that the work we do often carries sensitivity.  

Strong candidates may also have:

• Existing security clearances.  
• Fluency in, or eagerness to adopt, modern productivity tools (coding agents, Linear, Slack) and willingness to use AI tools as part of daily work.  

Selection process 

Applications are reviewed on a rolling basis. We encourage you to apply early. 

• Application - CV and a short cover note (max 500 words) explaining what draws you to this role and what you would bring. Please also indicate which team you would prefer to join.  

• Screening interview - Lightweight conversation on trajectory, motivation and fit.  

• Work test - A take-home exercise designed to test programme management judgement in context. You will receive this after the screening interview. 

• Technical interview - Discussion of your work test.