Owain Mogford

I partner with AI labs, from frontier model developers to early-stage startups, as an independent adversarial security researcher. My work involves finding universal jailbreaks against state-of-the-art models, the design of autonomous red-teaming systems that scale adversarial testing beyond manual effort, and applied research across the OWASP Top 10 for LLM applications. I specialise in finding the failure modes that matter before they reach production, and in building the tooling that makes that search systematic rather than ad hoc. As a member of Anthropic's private bug bounty programme, I work alongside some of the leading practitioners in the field and have made substantive contributions to real-world AI safety.

We're proudly a not for profit business. Our sole aim is to help make letting properties a better experience, for everyone.

I have ranked among the top 15 AI jailbreakers worldwide in Gray Swan Arena competitions, with thousands of unique successful jailbreaks to my name. This track record of consistent, high-volume adversarial work against live models has led directly to being headhunted for Senior AI Security roles.

I analyse and monitor the work of our third party contractors to ensure service delivery and contract compliance. In my role, I aggregate and analyse complex data from across the entirety of the council's services to surface opportunities to senior leaders and support data-led decision making.