At CISUC/LASI, I developed machine learning pipelines for the AI4Birth Project to predict mode of delivery from clinical tabular data and obstetric ultrasound images. I applied multimodal learning, representation learning, and evolutionary optimisation.
I evaluated predictive performance using retrospective data and prospective validation against clinical practice, analysing how models performed in a real-world healthcare setting. At ULS Coimbra, I investigated LLMs for extracting and interpreting clinical information from free-text notes, comparing automated and manual approaches.
My work includes research on AutoML embedding optimisation, multimodal prediction, and explainability for obstetric medical imaging. I also coordinated the organisation of the XX National Meeting of Biomedical Engineering Students, contributing to planning and execution with a team of 31 students.
