AI Security Research Scientist - Applied Mathematics
The Role We are seeking an AI Security Research Scientist to join The Alan Turing Institute s Defence & National Security team, helping to strengthen the robustness, reliability, and resilience of AI systems used in high stakes environments. This is an exciting opportunity for a mathematically strong researcher to apply advanced modelling, optimisation, statistical, numerical and experimental skills to cutting-edge AI security challenges, working closely with researchers, engineers and external partners to turn innovative ideas into practical capabilities. The role would suit someone with a PhD or equivalent experience in a highly quantitative field, strong programming skills and the ability to work from first principles, learn quickly and communicate complex research clearly. Prior experience in machine learning or AI security is welcome but not essential, and eligibility for Developed Vetting (DV) clearance is required. Your Profile We are looking for a highly analytical researcher with a strong grounding in applied mathematics or a similarly quantitative discipline, backed by a PhD or equivalent experience involving advanced algorithmic, statistical or numerical methods and computer programming. The ideal candidate will be fluent in one or more modern data science languages, able to communicate research clearly through technical outputs and confident working independently while collaborating effectively with others. They will be someone who learns quickly, thinks from first principles and is excited by the challenge of applying rigorous mathematical and experimental thinking to complex AI security problems. Experience in areas such as optimisation, robustness, verification, deep learning, cybersecurity or AI red teaming would be an advantage, but just as important is the ability to move comfortably between theory, modelling and experimentation, engage with technical and non-technical stakeholders and meet the eligibility requirements for Developed Vetting (DV) clearance. How You'll Make an Impact
- Adapt, assess, and extend existing research (from within the Institute and beyond) to understand how it behaves in realistic operational settings.
- Formulate research questions that make novel ideas testable, measurable, and actionable for partners.
- Design and run rigorous deep learning and AI security experiments - from robustness and adversarial evaluations to interpretability studies and stress-testing emerging methods.
- Generate insights : analyse results, identify failure modes, characterise risks, and evaluate whether a method is ready to progress toward real-world use.
- Prototype and iterate on new approaches, working closely with Research Engineers to ensure experiments are rigorous, reproducible, and well implemented.
- Translate findings into technical reports, presentations, demonstrations, and early-stage capabilities that inform partners operational decisions.
- Contribute to the wider research community , sharing knowledge internally, participating in discussions, and contributing where appropriate to publications and open-source outputs.
- Your current nationality
- Your nationality at birth
- Other nationality (include dual nationality if applicable)
- Confirmation that you have been residing in the UK for the past 5 years (if you haven t, please provide details of when and where you resided and the reason)
- Country where you were born.
- County in which you were born.
- Town where you were born.