Research Associate in Private and Trustworthy AI: Information-Theoretic Foundations
Imperial College London
In this Postdoctoral role, you will pursue information-theoretic questions for trustworthy generative models within INFORMED-AI, a UKRI AI Hub. You will shape your own research agenda, develop provable algorithms, and evaluate them on state-of-the-art models. The position offers collaboration across Imperial, Bristol, and the wider Hub, with opportunities to publish in top venues and mentor junior researchers. Your work will address watermarking, machine unlearning, and privacy-preserving learning to support responsible AI deployment. This is a full-time, in-person role based in London, funded for up to 24 months starting January 2027 or soon after, under Prof. Deniz Gunduz.
Pay / Benefits- sector-leading salary
- 39 days off per year
- generous pension schemes
- GPU computing resources
- funded research visits and conference travel
- opportunity to co-supervise PhD students
- Shape and lead your own research agenda within the specified themes
- Conduct original, internationally excellent research with theoretical and practical impact
- Develop theory with formal guarantees and translate it into practical algorithms for generative models
- Prepare reports for research sponsors and disseminate findings at conferences
- Publish in top information theory, ML, and security venues and collaborate across the Hub
- Collaborate with Bristol and Hub researchers and mentor PhD students
- Present results and contribute to INFORMED-AI Hub activities
- PhD in electrical engineering, computer science, mathematics, statistics, or related field
- Publications or track record in leading venues; practical research experience
- Strength in one or more: optimisation methods, information and coding theory, privacy and learning theory, deep learning design and training
- Programming in Python with PyTorch or JAX
- Desirable: prior work on watermarking, machine unlearning, differential privacy, membership inference or memorisation; hands-on experience with LLMs or diffusion models
- Knowledge in information theory, coding theory, learning theory, privacy, optimisation, and generative models
- ability to conduct literature reviews
- creative problem solving
- excellent verbal communication
- information theory
- coding theory
- privacy (differential privacy) and statistical learning theory
Reference: WJ-747_31300481 · Posted 20 Nov 2026, 12:46 AM
More jobs like this: