Published research · Applied AI evaluation
Evaluating AI on nuclear medicine knowledge tasks
Co-authored research comparing large language models on 600 nuclear medicine technology board examination–style questions.
600-question evaluation · Published May 2025
Context
Healthcare AI needs task-specific evaluation. I co-authored a comparison of large language models on nuclear medicine technology board examination–style questions, published in the Journal of Nuclear Medicine Technology.
My role and ownership
Co-authored with Michael A. Oumano. The publication, rather than this portfolio summary, is the source for the study methods, results, and author contribution details.
The work
- Compared model performance on a defined set of 600 board examination–style questions.
- Connected AI evaluation with a domain in which I have clinical and operational experience.
- Shared the work through publication so readers can examine the methods and limitations instead of relying on a product demo alone.
Outcomes
- Published research in the Journal of Nuclear Medicine Technology in May 2025.
- A concrete example of evaluating AI against a domain-specific knowledge task.