Clinical AI

Medication safety

SynRx

An explainable decision-support system that evaluates drug interactions, contraindications, and dose mismatches in context.

Clinical decision supportRAGFastAPIMedication safety
01 · Problem

Medication alerts are only useful when they are timely, specific, and grounded in the patient's clinical context. Generic rules create noise and erode clinician trust.

02 · My role

Co-founder and technical lead, responsible for clinical translation, system architecture, inference services, evaluation, deployment, and hospital integration.

03 · System

Python and FastAPI services, MySQL, RabbitMQ, Docker, and a retrieval-augmented evidence pipeline built from verified pharmacology sources.

04 · Scale

1.2M+ medication-safety alerts across a network serving 43 hospitals.

05 · Evidence

Traceable source material, curated clinical rules, validation datasets, and mixed-methods implementation evaluation using RE-AIM and CFIR.

06 · Outcome

Patient- and context-specific guidance embedded in prescribing workflows, with an auditable path from evidence to alert.

Next case study

Longitudinal laboratory intelligence