I am a data scientist who enjoys learning from people with different professional backgrounds. In my maritime projects, seafarers, port teams, and engineers have helped me understand aspects of vessel and port operations that I could not learn from operational records alone. They bring practical details that influence performance and decisions. I also explain how the model reaches its results and where its limits are, so they can question it before using it. This exchange has shaped my work in the maritime industry. I have built models that predict vessel performance and fuel use from historical operating data while accounting for changing weather and sea conditions. I have also examined vessel efficiency from a wider operational perspective, considering how technical conditions, working practices, and human decision-making interact to shape performance. Through this work, I identified where data and decision-support tools could improve operations while accounting for the constraints crews and operators work under. I have also developed port digital twins and energy models for decarbonisation. Port teams can use these tools to compare infrastructure and energy options and understand how each could affect energy demand and emissions.

In my PhD, I explored how AI could personalise maritime training and improve the learning experience for seafarers. I built a system in which AI agents assessed learner responses and provided feedback tailored to each person. After fine-tuning the model for maritime scenarios, I evaluated the system with seafarers to understand how they perceived the feedback and what shaped their experience of interacting with AI.

Beyond maritime operations, I have applied data science to problems in property technology and predictive medicine. At Tiko, I worked with housing and market records to build production pricing models for homes in Spain, turning property characteristics into estimates for individual properties. At BASIRA Lab, I worked with graph-structured brain-connectivity data from Alzheimer's disease and autism studies. I trained graph neural networks to classify diagnostic groups and examined which brain regions and connections had the strongest influence on their predictions. I was fascinated by how much we could learn from the same brain data by approaching it with different models. I began to see the same potential across industries, where organisations collect large amounts of data but use only a fraction of it. I wanted to find ways to bring those sources together and turn them into information people could use. These roles gave me experience with distinct data structures and modelling questions across commercial products and academic research.

My interests include applied data science in maritime, energy, infrastructure, and logistics, with a focus on complex operational data and physical systems.