About
Data Scientist in data privacy, security, and record linkage
Software Innovation Institute, Australian National University.
Profile
I am a Honorary Lecturer in the Software Innovation Institute at the Australian National University. My primary area of research is preserving the privacy of individuals through techniques including distributed secure data management, differential privacy, and federated learning.
I am currently involved in projects working with Solid and the concept of Personal Online Datastores ‐ personal data vaults that gurantees individuals true ownership of their own data and proper access management to it, rather than handing it over to third parties.
My long-standing research interest is in privacy-preserving record linkage (PPRL), particularly developing privacy attacks on PPRL techniques and investigating more secure techniques that are robust against existing vulnerabilities. I currently supervise two PhD students in the area of PPRL and Record Linkage and have supervised multiple undergraduate and master's students in the past.
More recently, I have been exploring the intersection of Personal Online Datastores and private data management with AI. More specifically, I am interested in how these technologies can be combined to create more secure data access and execution environments.
Outside research I am a keen photographer, studying it as a hobby. I enjoy travelling and photographing landscapes, cityscapes, woodlands, and wildlife.
Research interests
- Privacy-preserving record linkage
- Probabilistic record linkage
- Cryptanalysis PPRL techniques
- Data privacy and security
- Cryptography
- Differential privacy
- Federated learning
- Decentralised data architectures
- Solid personal online datastores
- Data mining
Current projects
Solid and Pod-based systems
Applications and open-source Dart and Flutter libraries for Solid authentication, decentralised identity, and the interfaces built on top of them. See: SII GitHub and Personal GitHub.
Custos
A personal context vault. An AI assistant receives only the facts a given task needs from a POD; decryption and access control happens in a broker the user controls.
Rio
A software suite supporting the secure collection, storage, and management of EEG data for a large-scale clinical trial. A colloboration with ANU School of Medicine and ANT Neuro, Germany.