DeepMetrics

Publications
  1. Overview of VentPilot development and validation: data collection at Seoul National University Hospital, development with intensivist feedback and offline reinforcement learning, and internal and external (Mayo Clinic) validation

    Development and validation of VentPilot: an AI-based recommendation system for mechanical ventilation

    Hong Yeul Lee, Gaon An, Yeonwoo Jeong, et al.
    Journal of Intensive Care, 2026

    In validation cohorts from Seoul National University Hospital and Mayo Clinic, fitted Q-evaluation estimated higher returns for VentPilot than for observed clinician behavior, and higher concordance with its recommendations was associated with more ventilator-free days within 28 days (mean differences of 5.0 and 3.1 days).

Contact Us
  • US
  • 222 East 44th Ave, New York, New York. 10017
  • Korea HQ
  • 205-ho, 396 Seocho-daero, Seocho-gu, Seoul, 06619, Republic of Korea
  • E-mail
  • careers@deepmetrics.co
  • contact@deepmetrics.co