Healthcare AI company nference and Cedars-Sinai Medical Center announced a long-term collaboration on November 20, 2026, to develop a federated artificial intelligence research platform that will analyze de-identified clinical data from more than 5 million patients spanning over a decade of longitudinal care. The planned infrastructure aims to enable researchers to build and validate AI models across therapeutic, diagnostic, and surgical domains without centralizing sensitive patient information.

Federated Architecture Addresses Privacy Concerns

The platform will employ a federated software design that allows AI models to be trained and evaluated on clinical data that remains within Cedars-Sinai's secure environment. According to the announcement, the system will incorporate physician adjudication to curate and harmonize disparate clinical data formats—a critical step given the heterogeneity of electronic health record systems and clinical documentation practices.

The data types planned for inclusion span the full spectrum of modern clinical care: de-identified clinical notes, laboratory results, medications, diagnostic reports, radiology and pathology images, genomic information, electrophysiology waveforms, vital signs, flowsheet measurements, and consented biospecimens. The platform is designed to receive frequent updates as routine care progresses, maintaining current longitudinal trajectories for enrolled patients.

Venky Soundararajan, founder and co-CEO of nference, and Sumeet Chugh, MD, vice dean and chief artificial intelligence health research officer at Cedars-Sinai, are leading the collaboration.

Intended Applications Across Research and Industry

The platform is being designed to support investigator-led studies, grant-funded research, education initiatives, quality improvement projects, and collaborations with life-science industry partners. Potential downstream applications identified in the announcement include drug discovery, clinical trial design, and future diagnostic development, though these represent intended use cases rather than achieved outcomes.

The federated approach positions the infrastructure as an alternative to centralized clinical data aggregation models. By keeping patient data local and bringing computational models to the data rather than moving data to centralized repositories, the collaboration aims to address growing regulatory and institutional concerns about health data sharing while still enabling large-scale AI research.

A Model for Academic Medical Centers

For academic medical centers, the partnership represents a potential template for converting routine clinical data into research assets while maintaining privacy standards and institutional data governance. The model could enable participation in multi-site federated research networks without requiring direct data sharing between institutions—a persistent barrier to collaborative clinical AI research.

The platform's success will ultimately depend on whether the infrastructure produces actionable research findings and validated AI models that translate into clinical benefit. The collaboration is still in development, with the announced capabilities representing planned features rather than operational systems. As regulatory scrutiny of health AI intensifies and competition among clinical data platforms grows, the federated architecture and physician-adjudicated data curation approach may offer nference a differentiated position in the clinical AI infrastructure market.

Sources

  1. Nference Plans Federated AI Research Platform with Cedars-Sinai — unite.ai