ºÚÁÏÍø´óʼÇ

Acute stroke is a leading cause of death and long-term disability, with treatment decisions that must be made within minutes. Clinicians face an overwhelming amount of information from multiple sources, such as brain scans, test results, bedside observations, and doctors’ notes, yet current decision support systems usually only process one or two types of data. As a result, clinicians are forced to piece together fragmented information under intense time pressure, and important details may be overlooked. This increases the risk of delays and uncertainty, directly affecting patient outcomes.

This PhD project will address this gap by developing an AI-driven system that can bring together structured data, free-text notes, imaging and temporal data. Clinicians will be able to receive concise, context-aware, and evidence-informed decision support. This will reduce cognitive burden, improve decision speed, and support personalised care, ultimately improving outcomes for people experiencing stroke.

Candidate requirements:

The ideal candidate will have a strong background in medical imaging and natural language processing (NLP), with hand-on experience in deep learning frameworks (e.g., MONAI and Pytorch), clinical NLP/large language models (LLMs), Retrieval-Augmented Generation (RAG) and multimodal fusion. Familiarity with medical imaging analysis, LLMs, and health data is essential to complete this PhD project, alongside solid statistics and careful data governance. Evidence of research potential (e.g., peer-reviewed publications) is expected. A demonstrated interest in AI in healthcare and the ability to collaborate in interdisciplinary settings will be highly valued.