My current research interest is integrating machine learning techniques into resource constrained devices.  Techniques include traditional approaches (decision trees, naive bayes, logistic regression) and nascent graph neural networks.  Current focus is on anomaly detection tasks for edge devices for cybersecurity applications.  I am also involved in pervasive health monitoring using wearables and smart home sensor devices.  Current application is determining human activity recognition for Parkinson's disease using multi-modal data sources as part of the TORUS project.

I was previously a Co-Investigator for the SYNERGIA Project to secure an end-to-end IoT platform.  I am currently a Co-Investigator on the ELABORATOR project that is designing and deploying specific innovations and interventions towards safe, inclusive and sustainable urban mobility for smart cities.  I am also a Co-Investigator on the CHARIOT Project that aims to mitigate adverse consequences of ransomware attacks in Industrial IoT (IIoT).

I previously worked on pervasive healthcare monitoring as part of the EurValve project at Bristol where I designed and implemented the "SPHERE in a box" kit.  My initial research was on efficient routing and broadcast approaches for IoT networks.

Other research includes adversarial machine learning for image classification/privacy, physical layer key agreement for IoT devices, and text classification for fan fiction content ratings.