The Hong Kong Polytechnic University is seeking a Research Assistant to join an RGC-funded project focused on developing a knowledge-driven multimodal framework for 3D MRI analysis and interpretation. The project aims to ground visual language models (VLMs) for diagnostic reasoning, integrating medical imaging, vision-language models, and explainable AI to build an end-to-end, clinically meaningful system. As a Research Assistant, you will work on developing a VLM-based MRI Q&A system that incorporates knowledge distillation, 3D segmentation, and radiomics grounding. The role involves aligning pixel-level signals with high-level reasoning through zero-shot and fine-tuning approaches. You will design grounding components for precise 3D segmentation and radiomics, integrate knowledge graphs and causal reasoning for explainable medical Q&A, and validate both public and proprietary datasets to improve segmentation, grounding, and reasoning. The position offers the opportunity to deliver interpretable, clinically actionable insights and contribute to broader research tasks as needed. Applicants should have a bachelor's degree (or equivalent), preferably a master's degree, in computer science, artificial intelligence, machine learning, or related fields. Strong foundations in probability, linear algebra, and optimization are required, along with experience in medical image analysis. Hands-on experience with knowledge distillation, vision-language models, and causal or structured reasoning is essential. Proficiency in Python and modern machine learning frameworks, familiarity with imaging libraries, a demonstrated publication record, strong data curation and validation experience, and excellent written and spoken English are also required. This position offers impactful research opportunities that connect MRI signals to reasoning for clinical decision support, growth in 3D medical imaging and multimodal explainable AI, and close mentorship with an active research team. The role is funded by the Research Grants Council (RGC), and successful candidates will have the chance to contribute to publications and prototypes. For more information or to apply, contact Dr. Sheheryar Khan at [email protected].