Edinburgh, Scotland, United Kingdom
Machine learning researcher and software engineer with a background spanning academic NLP and scalable commercial systems. Over the past decade, my work has moved from building high-throughput risk architectures in financial services to leading applied ML initiatives, most recently at Rithum. I focus on bridging complex engineering requirements with practical machine learning, taking projects from making significant architectural proposals through to delivering end-to-end production pipelines. In recent commercial roles, I have led the development of robust AI solutions utilising generative coding and agent-based workflows. This involves building architectures for multimodal classification, structured knowledge extraction, and forecasting using causal models such as Double Machine Learning. Beyond the core technical implementation, I place a strong emphasis on team development and sharing insights. I regularly mentor colleagues, deliver research talks on emerging methodologies, and enjoy participating in internal hackathons, having recently won prizes for prototyping novel reasoning and data pipelines. Academically, I hold a PhD in Informatics from the University of Edinburgh. My research investigates computational narrative, focusing on how systems infer salience, suspense, and surprise through self-supervised learning, memory, and knowledge bases. This intersection of reasoning and memory continues to drive my open-source work. My current project, Shadow-Loom, explores neural-symbolic graphical approaches that combine semantic memory graphs with deep learning to better comprehend and generate complex structural narratives.