DrSoteris Demetriou

Dr Soteris Demetriou is an Associate Professor in the Department of Computing at Imperial College London, where he leads the Applications, Platforms, and Systems Security (APSS) research group and serves as Director of Imperial's Academic Centre of Excellence in Cyber Security Research. He received his PhD and MSc in Computer Science from the University of Illinois at Urbana-Champaign (USA) and a Diploma in Electrical and Computer Engineering from the University of Patras (Greece).

Dr Demetriou's research asks how we build AI-integrated systems that are not just functional but trustworthy; secure against adversarial manipulation, private by design, and accountable to the norms governing their deployment. By analysing operating systems, networking protocols, machine learning models, and side-channels, his work has uncovered design flaws leading to  severe exploits and privacy leakages on the Android OS, Amazon services, commodity IoT devices, and LiDAR-based 3D object detectors, affecting millions of users. His recent focus centres on the safety and accountability of generative AI: studying attribute leakage and unsafe content generation in text-guided diffusion models, and developing neuro-symbolic methods for detecting, explaining, and enforcing policy compliance in large language and diffusion models.

Dr Demetriou's research has been published extensively at leading venues in security, systems, and applied machine learning, including NDSS, ACM CCS, USENIX OSDI, ACM SOSP, ACM MobiSys, IEEE INFOCOM,  Interspeech, and IROS. His work has received multiple distinctions, including a Distinguished Paper Award at NDSS, and has influenced practice through technology transfer and awarded US patents. He has fostered collaborations spanning more than 100 researchers across more than 20 institutions in academia and industry, including a visiting appointment at Meta working with several teams on hyperscale distributed computing, and is an active member of the security research community.

You can read more about this work at his research group's website: apss.doc.ic.ac.uk.


--ANNOUNCEMENTS--

We are always looking for excellent PhD candidates to join our growing team. If you are passionate about solving security, safety, and privacy problems in systems and machine learning models, have strong evidence of research experience (demonstrated through publications in international venues), and you are interested in joining APSS, please send me an email using “Prospective Security PhD Student” as subject.