Partnerships
Nference Plans Federated AI Research Platform with Cedars-Sinai

Key PointsThe platform is planned infrastructure rather than a delivered product with…
View more
- The platform is planned infrastructure rather than a delivered product with capabilities still described as intended.
- Federated software and physician adjudication aim to protect privacy while harmonizing messy clinical data formats.
- Its real test is whether curated longitudinal data helps investigators validate hypotheses and build useful models.
nference announced a long-term collaboration with Cedars-Sinai on November 20, 2026, to build a secure platform for analyzing de-identified clinical data from multiple areas of patient care. The planned research foundation includes more than a decade of records from over 5 million patients, following their care over time.
The platform is still to be developed, with its research and AI functions described as intended capabilities. The work is designed to convert information generated during clinical care into research-ready data across therapeutic, diagnostic and surgical areas, giving investigators a common basis for analyzing records across disparate and unstructured formats.
Privacy, Curation and Data Scope
nference will architect the system with federated software designed to protect patient privacy and harmonize data from different clinical formats. Longitudinal curation, the organization of records as patient care unfolds, will include physician adjudication, bringing doctor review into the process of preparing data for research.
The de-identified inputs will include clinical notes, laboratory results, medications and diagnostic reports, alongside radiology and pathology images, genomic information, electrophysiology waveforms, vital signs and flowsheet measurements. Biospecimens collected with patient consent will also be included.
Frequent data refreshes will add new information as routine care progresses. nference envisions this continually updated resource helping researchers gain finer-grained insight into how diseases begin and progress, an intended benefit rather than a demonstrated research outcome.
Studies, AI Models and Scientific Objectives
The platform is intended to support investigator-led studies, grant-funded research, education and quality improvement, as well as collaborations with life-science industry partners. It will also provide a foundation for Cedars-Sinai researchers and affiliated innovators to develop and evaluate new AI models.
Drug discovery, clinical-trial design and future diagnostics are potential downstream uses identified by the companies. Those applications depend on the studies and model evaluations the platform is meant to enable; they are not outcomes announced as already achieved.
Venky Soundararajan, founder and co-CEO of nference, described the planned foundation as whole-body spatial intelligence, intended to triangulate disparate and unstructured modalities so difficult scientific questions can become more tractable. “Understanding health and disease requires connecting what is happening across organs, across clinical modalities, and across decades of each patient’s medical journey,” he said.
Sumeet Chugh, vice dean and chief artificial intelligence health research officer at Cedars-Sinai, said the institution expects the collaboration to help clinicians and scientists learn from collective patient-care experience, validate scientific hypotheses and accelerate biomedical discovery.
The collaboration’s immediate deliverable is research infrastructure: a privacy-focused, regularly refreshed environment that joins clinical curation with AI model development and evaluation. Its broader significance will depend on whether those combined data and curation mechanisms help investigators validate hypotheses and produce useful models across the many kinds of patient information the platform is designed to connect.












