Projects
Beyond LLMs: Advancing Clinical Intelligence
CHMI researchers are pioneering a new foundation model that goes beyond today’s LLMs. As CHMI advances the science, its first spinout, Knit Health, is translating this work into tools that deliver reliable, real-time clinical intelligence at scale.
Predicting Sudden Cardiac Death and Rethinking Care Delivery
Ziad Obermeyer is using ECG waveform data to predict who is truly at risk of sudden cardiac arrest - identifying patients who need intensive care and those receiving unnecessary, costly interventions. This project is part of CHMI’s broader effort to harness new kinds of clinical data and develop portable, efficient, and scalable tools that could transform how healthcare is delivered over the next decade.
California Medi-Cal’s new Risk Stratification Algorithm
In partnership with the State of California, Maya Petersen and Jonathan Kolstad are developing a statewide algorithm that redefines how Medicaid patients are segmented by risk, optimizing care for millions while improving resource use.