Speaker Profile
Ph.D., AI/ML, Apple
Biography
Mohsen Hejrati works on building medical AI systems that are safe enough to be trusted with real patients. His current focus is on the hard problems that surface when AI agents move from demos into clinical settings: keeping models grounded in evidence rather than plausible-sounding output, knowing when a system should defer to a physician instead of acting, evaluating safety on the rare and high-stakes cases that averages hide, and earning the trust of clinicians who have to live with the consequences. He is interested in how these systems fit into existing clinical workflows without adding risk or burden.
Earlier he was Senior Director of AI at Genentech, where he led AI teams and co-architected clinic-in-the-loop approaches to drug development. Before that he worked on machine learning at Google X. He is a repeat founder with three prior exits, and his research background is in machine learning.
Session Abstract – PMWC 2027 Silicon Valley
Track Chair:
Vincent Liu, Kaiser Permanente
PMWC Award Ceremony
• Ziad Obermeyer, UC Berkeley (Luminary)
Featured Speakers
• Ziad Obermeyer, UC Berkeley
• John Halamka, Mayo Clinic Platform
• Ashley Beecy, Sutter Health
• Shiv Rao, Abridge
• Munjal Shah, Hippocratic AI
• Vivek Natarajan, Google Health
• Vincent Liu, Kaiser Permanente
• Mark Sendak, Duke Health
• Joshua Denny, NIH
• Chris Lunt, Critical Path Institute
• Ida Sim, UCSF
• Travis Zack, OpenEvidence
• Danton Char, Stanford
• Brian Anderson, CHAI
• Julia Adler-Milstein, UCSF
The Year in Clinical AI
What shipped, what failed, and what should change next.
AI Agents in Clinical Workflows
Ambient documentation, patient-facing agents, and operational agents in real care.
Foundation Models in Healthcare
What makes a clinical foundation model useful, deployable, and trusted.
AI, Human Cohorts, Trials & RWE
Turning clinical data into stronger trials and decision-grade evidence.
AI Safety, Governance & Evaluation
Avoiding harm at scale when algorithms move into production.
AI + Pharmacogenomics
How decision support and drug-response data can make each other more useful.
PMWC Hall of Impact
Previous Speakers Include
Nobel laureates, technology founders, regulators, CEOs and scientific pioneers who have taken the PMWC stage.
