About this role
Please click "Apply Now" to respond to the initial questions in the application form and submit your CV / Resume and cover letter. We ask that you spend no more than 2 hours preparing your responses to the prompts. Application Deadline: Sunday, August 30, 2026, at the end of the day (23:59) Anywhere on Earth (UTC-12:00). About the Role Our staff's time is one of RP's tightest bottlenecks. Over the last year, AI tools have crossed a threshold where they can meaningfully accelerate serious knowledge and administrative work. But they're also uneven, and it's easy to feel faster while producing plausible-sounding work that doesn't hold up. As an organisation that cares about getting things right, we don't want to adopt AI on vibes. We want to know, rigorously, where it genuinely improves our work and where it doesn't. That's what this role is for. You'll lead RP's effort to figure out — empirically — which of our research and operational tasks AI can reliably and usefully augment, build and deploy the tools and workflows that prove out, and steer us away from the ones that don't. This is as much a measurement and research problem as an adoption one, which is why we want someone who understands both research rigor and the real capabilities and limits of modern AI tooling. You'll be the lead for this area at RP and will report to the Principal Research Director, David Moss. The role may be offered at a more junior or more senior level (and titled accordingly) depending on the successful candidate's experience Hours and term: Our default is full-time on an initial 12-month contract with the possibility of extension . We're open to part-time for the right person, and open to a shorter initial term with the option to re-engage afterwards. Tell us what arrangement works for you. Contact: Please email [email protected] if you have any questions. Own RP's strategy for using AI in our work. Propose and prioritise where to invest, set the direction, and be the directly responsible individual for how AI is used across the organisation. Design and run rigorous, decision-relevant evaluations. Test AI tools and workflows against the work we actually do, measuring real effects on quality, speed, and reliability, not just adoption or self-reported time savings. Distinguish clearly what AI can be trusted to do from what it can't. Build and prototype. Get hands-on with agentic tooling (e.g. Claude Code), plugins, MCP integrations, and lightweight internal tools — first to test approaches, then to scale the ones that work. Partner with teams across RP. Sit with researchers and operational staff to understand their workflows in detail, surface high-value use cases, and embed what works into day-to-day practice. Turn findings into guidance. Translate evaluation results into clear, current advice, policy, and training, and raise the organisation's overall AI fluency. Assess safety and security of AI tools. Identify risks from AI tool use, including cybersecurity, data protection and legal compliance, and document clear recommendations for handling them. Stay ahead of a fast-moving field and keep RP's approach evidence-led rather than hype-led.