Inspiration
Heart disease kills silently. Millions of people are told they are “high risk,” yet never understand why—or what they can realistically do about it. Most tools stop at a number, leaving users confused and powerless.
CardioInsight AI was built to change that narrative by turning cardiovascular risk prediction into something understandable, transparent, and actionable.
What It Does
CardioInsight AI predicts a person’s 10-year heart disease risk and enhances traditional models by:
Explaining which factors drive the prediction
Showing how risk changes when health parameters improve
Offering guidance instead of just a probability
Working even when full medical data is unavailable
The platform supports both population-level analysis and interactive individual risk exploration.
How We Built It
We implemented a two-model system using XGBoost:
A clinical-grade model for high-accuracy predictions
A lightweight model for real-world accessibility
SHAP was used for explainability, what-if simulations for actionability, and Streamlit for deployment.
Challenges
Balancing recall for high-risk patients, ensuring calibrated predictions, simplifying explanations, and maintaining medical realism were key challenges throughout development.
Accomplishments
We delivered an explainable, user-focused AI system that goes beyond prediction—helping users understand and act on their risk.
Learnings
Trust, interpretability, and real-world usability matter just as much as model performance in healthcare AI.
What’s Next
Clinical validation, personalization, wearable integration, and fairness-aware modeling.
Log in or sign up for Devpost to join the conversation.