There is an urgent need to strengthen the HIV clinical workforce in the United States, as demand for HIV prevention and treatment continues to rise while the supply of trained providers declines. Contributing factors include an aging clinician population, limited HIV-specific training during medical education, and systemic disincentives to managing complex HIV care. Concurrently, medical education is evolving toward online and self-directed formats, but most continuing medical education (CME) electronic curricula remain passive and do not adequately develop clinical reasoning. Case-based learning (CBL) and flipped-classroom (FC) models are more effective but remain resource-intensive and difficult to scale due to reliance on faculty time. Artificial intelligence (AI) offers a promising solution to this CME implementation gap by enabling dynamic, interactive, and scalable learning experiences that simulate expert clinical instruction. We have developed and piloted PRISM (Precision Review and Interactive Simulation in Medicine), a large language model (LLM)-based educational tool that delivers simulated case discussions, diagnostic feedback, and content review aligned with the Johns Hopkins Infectious Diseases curricula. In early testing, PRISM demonstrated high engagement and usability among medical students. This project extends this work and proposes to evaluate and refine PRISM- HIV, an AI-driven, case-based learning platform aligned with an existing HIV e-learning curriculum (Foundations in HIV Medicine). Target learners include medical residents, general internists, and advanced practice providers, who need to build confidence and competency in HIV medicine to address the workforce crisis. PRISM-HIV simulates a faculty preceptor to guide learners through interactive HIV case scenarios, promote clinical reasoning, deliver precision review, and assess user responses. Case-based learning and scenario details will map to the e-learning curriculum and core competencies, and facilitate efficient learning needed for continuing medical education of post-graduate learners in a manner aligned with adult learning theory. We will conduct a three-part study to: (1) evaluate the accuracy and consistency of AI-generated HIV cases and responses; (2) assess acceptability, usability, and user experience among internal medicine residents, primary care providers, and HIV clinicians using mixed methods; and (3) conduct a hybrid implementation-effectiveness trial comparing PRISM-HIV-enhanced learning to lecture-based learning alone, with outcomes based on knowledge acquisition, self-confidence, and RE-AIM implementation metrics. If successful, PRISM-HIV will offer a scalable and flexible continuing medical education (CME) tool to build HIV clinical capacity with minimal faculty burden. This project has the potential to modernize HIV education, support ongoing workforce development, and serve as a model for leveraging AI to improve healthcare delivery across multiple clinical domains.