A clinical decision support system that listens to patient presentations, reasons through differential diagnoses, recommends appropriate tests, and surfaces when to refer — without ever diagnosing or prescribing on its own. Built on multi-layer NLP inference, knowledge graphs of conditions and symptoms, and an evaluation-first development process that treats under-triage as a hard failure.
Capabilities applied
Expert Systems Safety-Gated AI Knowledge Graphs NLP Inference
Technology
Python FastAPI YAML knowledge bases RL supervisor loop Evaluation harness
Interested in similar work?
Start a conversation