OptiSense AI

Built for outcomes, not demos

An oculomics platform that turns a standard retinal (fundus) photograph into clinical signal — using the eye as a non-invasive window into both ocular and whole-body health. Its production capability today grades diabetic retinopathy on the 5-class international clinical scale with EfficientNet transfer learning, retina-specific preprocessing, an ordinal head tuned for sensitivity, and Grad-CAM explainability. The same imaging-and-deep-learning foundation is built to extend to broader retinal biomarkers of cardiometabolic and neurological risk — all delivered as clinician-led screening and decision support.

Deployed on FEME's secure, cloud-native platform, it integrates with your existing systems, keeps humans in control of critical decisions, and scales from a single workflow to enterprise-wide operations.