AI-Driven Virtual Teaching Assistant (ViTA) to Advance Interdisciplinary Graduate Training in Medical Biophysics

Graduate education in emerging interdisciplinary fields such as medical biophysics (MBIO) faces a fundamental challenge: students enter with uneven preparation across quantitative, computational, and life science domains, leading to cognitive overload, burnout, and attrition. We present ViTA, an AI-driven Virtual Teaching Assistant, to be developed under the NSF Innovations in Graduate Education (IGE) program to enhance learning in interdisciplinary MBIO training. ViTA integrates discipline-aware large language models with retrieval-augmented generation to provide personalized learning support across MBIO core areas, including molecular biophysics, tissue mechanics, imaging physics, computational modeling, and statistical data analysis. Embedded in the Canvas learning platform, ViTA diagnoses learning gaps, adapts instruction, and generates faculty-facing analytics to guide curriculum refinement. ViTA establishes a scalable model for human–AI co-mentoring in graduate education and will be disseminated as an open framework to partner institutions, including HBCUs, to promote national STEM workforce development.