- PhD Student
I am a Computer Science PhD student at Downing College working under the supervision of Professor Robert Mullins and Professor Ross Anderson. My research is being supported by the Harding Distinguished Postgraduate Scholars Programme.
Biography
I matriculated at Churchill College in 2018, receiving a BA (Hons) in Computer Science in 2021. Following this, I completed an MSc in Advanced Computer Science at St Cross College at the University of Oxford in 2022. Prior to university, I attended St Paul's School which endowed me with the academic drive I now strive to maintain.
Research
I'm interested primarily in security as a broad academic and practical discipline. Specifically, I enjoy working with adversarial machine learning, real-world security, and more recently - AI safety. My secondary interests include economics, law, data science, and fintech.
Please reach out if our interests align - I'm happy to supervise Part II, III, and MPhil students for their dissertations/projects.
Teaching
I supervise students from a number of different colleges - Churchill, Magdalene, Homerton, Hughes Hall, Lucy Cavendish, St Catharine's, Sidney Sussex, Newnham, Downing, St Edmund's, and Murray Edwards. That is, 11 out of the 29 colleges that admit undergraduates.
I supervise the following undergraduate Computer Science courses:
- Part IA: Machine Learning and Real-world Data [20]; Software and Security Engineering [17]
- Part IB: Data Science [46]; Economics, Law and Ethics [42]; Security [23]
- Part II: Bioinformatics [9]
The respective number of students I taught for each course is shown in square brackets next to the course name. This brings the total number of groups I supervised to 67 over 8 terms, resulting in 219 hours of teaching so far. Those groups were made up of 157 students in total if we're double-counting, and 102 if we're only talking about unique individuals. Numbers are accurate as of Lent Term 2023.
I have supervised the following students for their Part II projects, in no particular order:
- 2023/24
- Dylan Mankin [CC]: Machine Learning for Financial Portfolio Management
- Ria Mundhra [DOW]: Developing a Graphical User Interface for Econometric Analysis
- Kai Tanna-Shah [CHU]: Creating an Interpretable Chess Engine
- Tayo Oluranti-Ahmed [DOW]: Forecasting Implied Volatility with Machine Learning Techniques
- Peter Isaksen [CHU]: Implementing and Evaluating an Adversarial Attack on an Aligned Large Language Model
I have also supervised the following external students:
- Micha Nowak as an academic advisor for his Master's Thesis in Artificial Intelligence at the University of Wuerzburg in Germany. This was done under the supervision of Professor Frank Puppe, and further advised by Dr. Adrian Krenzer.
Professional Activities
In my spare time, I run a video game porting and publishing company called mazette!
Publications
- Human-Producible Adversarial Examples [2023]
- David Khachaturov, Yue Gao, Ilia Shumailov, Robert Mullins, Ross Anderson, Kassem Fawaz
- Portable, Low-Cost, and Low-Power mmWave Radar Imaging for Concealed Weapon Detection [2022]
- MSc Dissertation, supervised by Professor Kasper Rasmussen
- IP patented and registered with Oxford University Innovation Limited
- Susceptibility of Hierarchical Reinforcement Learning to Adversarial Examples [2021]
- Computer Science Tripos Part II Dissertation, supervised by Ilia Shumailov and Dmitry Kazhdan
- Markpainting: Adversarial Machine Learning meets Inpainting [2021]
- David Khachaturov, Ilia Shumailov, Yiren Zhao, Nicolas Papernot, Ross Anderson
- Proceedings of the 38th International Conference on Machine Learning (Acceptance rate ~21.5%)
