Dyania Health Uses On-Premises AI to Extract Clinical Signals From Patient Records

Synapsis AI processes unstructured hospital records on-site to surface evidence for clinician review.

IT Solution Journal News Desk

Editorial illustration: A polished editorial illustration of a large glass bell jar on a dark pedestal containing a tall stack of worn, yellowed paper documents, from which luminous blue electrical filaments rise and converge toward a glowing amber point of light at the top of the dome, all set against a deep blue-black background.

AI-generated editorial illustration.

Dyania Health, a New Jersey clinical AI company, has built Synapsis AI, a clinical reasoning engine that processes physician notes, lab results and imaging reports within a hospital's own firewall using NVIDIA-accelerated GPUs. Models are built with NVIDIA NeMo and optimized via TensorRT for on-premises deployment, keeping patient data local. Clinicians review all outputs before acting. Cleveland Clinic applied the platform to a retrospective study of more than 7,800 aortic valve replacement patients spanning 2005 to 2021, with Dyania claiming the AI completed chart review that would otherwise have taken years. Additional use cases include clinical trial matching and care-gap identification.

Source: NVIDIA News. Original announcement: November 13, 2026.