Dyania Health Runs Synapsis AI on Hospital-Owned GPU Systems - Unite.AI

The deployment uses NVIDIA AI Enterprise across the software stack. Dyania trains its proprietary clinical models with NeMo, then uses TensorRT and TensorRT-LLM to optimize inference on hospital GPU servers. Inference therefore stays within the hospital's private network instead of sending patient records to an external environment.

NVIDIA emphasized that the system remains assistive: it surfaces evidence from medical records that clinicians must confirm and act on. That distinction matters because the deployment automates chart review and signal detection, not the clinical decisions that follow.

Clinical Reasoning Engine

Dyania Health describes Synapsis AI as a clinical reasoning engine that reads years of notes, operative reports, imaging reports and laboratory results, then answers clinical questions with evidence linked to the chart. Its outputs are clinical signals intended to show care teams which patients need attention and why; clinicians review the evidence and make every decision.

Cleveland Clinic Valve Study

The study examined patients who received aortic valve replacements between 2005 and 2021, reviewing up to 16 years of care per patient. Synapsis AI answered 88 clinical questions for each patient covering valve characteristics, co-existing conditions, laboratory trends, medications, follow-up ultrasound measurements and subsequent clinical events.

NVIDIA said the platform completed chart review across the full cohort, work that would have taken clinicians years to perform manually. Dyania reported that implanted valve type was the strongest determinant of how quickly a valve deteriorated, while age and other medical conditions influenced whether and when a second procedure occurred.

Future Deployment Plans

Dyania and NVIDIA plan to pursue health-system deployments in the United States and internationally, publish joint case studies and integrate Synapsis AI with NVIDIA's healthcare software stack. Technologies under evaluation include NVIDIA Confidential Computing for training and inference, NVIDIA Dynamo for distributed inference serving and Multi-Instance GPU for partitioning GPUs across concurrent clinical workloads.

The companies are also evaluating scaling to multiple Blackwell Ultra GPUs as processing moves from batch reviews toward near-real-time clinical insight at the point of care. Dyania says the tool is intended to inform care and requires external clinical validation before broader clinical reliance.