AI Healthcare / Research / 2026
TriageAI / ER-Predict
An AI-powered healthcare concept designed to address emergency department overcrowding by connecting real-time patient severity assessment with predictive capacity management. The concept combines TriageAI, which prioritizes patients and predicts short-term deterioration risk, with ER-Predict, which forecasts patient arrivals and resource demand up to 24 hours ahead, creating a proactive approach to emergency care management.
- Role
- Researcher / Product & Concept Designer
- Stack
- AINatural Language ProcessingMachine LearningTemporal Fusion TransformerProduct DesignHealthcare Systems
Problem
Emergency departments frequently experience overcrowding during health crises, seasonal outbreaks, and mass casualty events, while traditional triage remains heavily dependent on manual assessment and hospital resource planning tends to be reactive. This creates a risk of delayed care for high-acuity patients and inefficient allocation of beds, staff, and critical supplies.
Process
As a Researcher and Product & Concept Designer, I explored the intersection of emergency care, clinical AI, and predictive analytics to identify opportunities for a more proactive emergency department workflow. I researched existing triage and forecasting approaches, mapped the relationship between patient-level clinical risk and department-level capacity, and translated these insights into a two-module product concept. I contributed to defining the system architecture, AI workflow, safety mechanisms, data strategy, and evaluation metrics while ensuring that AI remained a decision-support tool with clinicians in the loop.
Solution
We proposed TriageAI / ER-Predict, an integrated AI system consisting of two complementary modules. TriageAI uses patient-reported symptoms, vital signs, and medical history to estimate urgency and short-term deterioration risk, while ER-Predict forecasts patient arrivals and downstream resource demand over a 24-hour horizon using historical trends, weather, calendar, and regional health data. The two modules are connected through a hospital dashboard to support both clinical prioritization and proactive capacity planning.
Result
The concept establishes a framework for shifting emergency department management from reactive response toward predictive and proactive decision-making. The proposed system targets improved triage accuracy, reduced waiting times for high-risk patients, and fewer severe resource shortages, while incorporating privacy-by-design and human-in-the-loop safety mechanisms. Through this project, I strengthened my ability to translate complex research and AI capabilities into a structured product concept that addresses a high-impact real-world problem.
Gallery
