Research

This scoping review synthesises evidence on AI supported conversational agents, including chatbots and digital companions, designed to support perinatal mental health among mothers, fathers, parents, and couples. Using PRISMA ScR and JBI frameworks, the study examines how these systems are designed, implemented, and evaluated, with a focus on computational methods, ethical governance, privacy preserving mechanisms, and inclusivity. The work identifies current practices and gaps, and provides guidance for the development of secure, equitable, and clinically meaningful AI driven perinatal mental health interventions.

This study proposes an improved five-level NNPC inverter leg that reduces switch voltage stress, with only one device rated at half of the input voltage. Experimental results demonstrate reduced losses, lower cost implications, stable capacitor voltage balancing, and competitive performance compared with existing NNPC topologies.

This work presents a self-balanced single-phase nine-level switched-capacitor inverter using a single DC source and a simplified control scheme. The topology achieves a four-times voltage boost with high efficiency and strong active and reactive power performance, validated on a 3 kW prototype.

This study analyses the impact of stator winding chording on the dynamic and steady-state performance of a five-phase synchronous reluctance motor. Analytical phase-variable models and finite element simulations show close agreement, with improved performance observed as the chording angle increases.

I had the opportunity to contribute to this systematic review in evolutionary biology, bringing computational and AI approaches to the study of how venom varies within a single snake species. I also developed the analytical framework used to synthesise the genetic, regulatory and functional evidence.