Research Interests
Throughout my career, I have had the privilege of working alongside clinicians in many hospitals to support patient care, participating in numerous clinical trials, working in industry at Philips and IBM, and contributing to non-profit research initiatives. This diverse experience shapes my translational approach to research.
Theranostics
From new ideas in data acquisition to patient satisfaction, theranostics carries many opportunities to improve patient care using AI. We work on multiple fronts related to improved imaging of therapy, exploring how artificial intelligence can enhance both diagnostic accuracy and therapeutic outcomes in nuclear medicine.
Patient Health Timelines & Foundation Models
Developing patient health timeline models, including ETHOS, for personalized decision support in the context of 4P medicine (Predictive, Preventive, Personalized, and Participatory). We study quantitative world models of clinical trajectories and how foundation models can support individualized risk estimation and clinical decision-making.
Related Publications:
- Beyond language: generative artificial intelligence as a general computing model for medicine (The Lancet Digital Health, 2026)
- MEDS — An Emerging Data Standard and Ecosystem for Health AI Research (NEJM AI, 2026)
- Foundation model of electronic medical records for adaptive risk estimation (GigaScience, 2025)
- Zero shot health trajectory prediction using transformer (npj Digital Medicine, 2024)
Towards Total AI in Radiology
Working towards comprehensive AI integration across the entire radiology workflow, from image acquisition and reconstruction to interpretation and reporting. This initiative aims to create seamless AI-assisted systems that enhance radiologist efficiency while maintaining and improving diagnostic accuracy.
Related Publications:
- Mamba Goes HoME: Hierarchical Soft Mixture-of-Experts for 3D Medical Image Segmentation (NeurIPS 2025)
- TabMixer: Noninvasive Estimation of the Mean Pulmonary Artery Pressure via Imaging and Tabular Data Mixing (MICCAI) [Code]
Improving Outcomes in Obstetrics
Applying AI and advanced imaging techniques to improve maternal and fetal health outcomes. Our work focuses on developing predictive models and decision support tools that can help clinicians make better-informed decisions in prenatal care and delivery.
Bayesian Statistics
Promoting Bayesian approaches to avoid many cognitive biases in data handling and clinical decision-making. This framework provides rigorous methods for making decisions under uncertainty, incorporating prior knowledge with new evidence to support more rational and transparent clinical reasoning.
Related Publications:
Other Professional Interests
I have worked on and contributed to various other domains:
- Imaging Systems Design — Design and optimization of nuclear medicine imaging systems
- Tomographic Image Reconstruction — Advanced algorithms for reconstructing images from projection data
- Quantitative Finance — Mathematical modeling and machine learning for financial markets and trading strategies
- Drug Discovery — Computational methods for accelerating drug discovery and therapeutic target identification
- AI Assistance in Surgery — Real-time AI tools to assist surgeons during procedures
- Data Visualization — Innovative visual representations of data
- Virtual Reality — VR applications in medicine, training, and immersive data exploration
For a complete list of publications, please visit my Google Scholar profile or ORCID.