Curriculum Vitae
Yu Akagi, MD
Physician · Researcher · AI Engineer
Physician with over seven years of clinical experience, and an AI researcher working on generative models for healthcare. Four years building and training large-scale generative models on real-world EHR data, with award-winning research presented through the American Medical Informatics Association (AMIA). Currently leading development of an EHR simulation environment for training, testing, and evaluating AI agents that operate electronic health records autonomously. Deep working knowledge of EHR systems, and extensive collaboration with healthcare stakeholders in Japan and the United States.
Experience
Jun 2026 — Sep 2026
Lawrence Livermore National Laboratory
InternshipCalifornia, United States
- Data Science Summer Institute (DSSI)
Dec 2025 — Present
National Institute of Informatics
Healthcare AI ResearcherJapan
Apr 2023 — Mar 2027
The University of Tokyo
Healthcare AI ResearcherJapan
- Clinical workflow automation with agentic AI, focused on autonomous operation of the electronic health record.
- Generative foundation models trained on large-scale real-world clinical data.
Apr 2020 — Mar 2023
Japanese Red Cross Medical Center
Clinical Fellow, Infectious DiseasesTokyo, Japan
- Rotations: Kanto Rosai Hospital, Division of Infectious Diseases (Oct — Dec 2021); Shizuoka Cancer Center (Apr — Jun 2022).
- Awarded Best Fellow among all trainees.
Apr 2018 — Mar 2020
Japanese Red Cross Medical Center
Resident PhysicianTokyo, Japan
Education
Apr 2023 — Mar 2027
The University of Tokyo
PhD in Medical InformaticsJapan
Apr 2012 — Mar 2018
Yokohama City University
Medical Doctor (MD)Japan
- International clinical training at the University of Oxford (UK) and Université de Paris (France), 2017.
Awards
2025
KDDM Innovation Award — First Place
American Medical Informatics Association (AMIA)
- AMIA Annual Symposium 2025
2025
8th Academic Paper Award
Japanese Association for Medical Informatics (JAMI)
- Second place for the best paper of the year.
Certifications
2023
Board Certified in Internal Medicine
Japan
2023
ECFMG Certification
United States
2018
Japanese Medical License
Japan
Research
Selected research
TwinEHR — digital-twin EHR platform
- In-silico simulation of patient trajectories, inside a standard EHR interface, driven by the Watcher model.
Watcher — generative EHR model
- Pretrained on over 200 million real clinical records. Simulates detailed patient timelines across admissions, diagnoses, medications, and laboratory tests.