Case Study · 2025
HealthHive — Health Monitoring & Risk Prediction
iOS health vitals tracking application connected to a Flask ML backend providing real-time cardiovascular risk prediction with 95% model accuracy.
Context
Preventative healthcare relies on continuous vital signs tracking and proactive risk scoring. HealthHive was built to give users personal health tracking coupled with predictive risk modeling.
The Challenge
Most vital tracking apps act merely as passive logs without actionable risk intelligence. Connecting an intuitive mobile interface with a validated predictive model required a robust multi-tier architecture.
Engineering Process
Predictive ML Model Training & Validation
Trained and evaluated regression models on cardiovascular clinical datasets, applying feature engineering on blood pressure, resting heart rate, glucose levels, and SpO2 to reach 95% predictive accuracy.
Flask ML Serving API
Engineered a lightweight Flask microservice exposing REST endpoints to consume live user vitals, run inference, and return instant risk probability scores and category classifications.
Native iOS Interface in SwiftUI
Crafted clean iOS dashboards with visual trend charts, vital entry forms, medication reminders, and instant risk indicator badges.
Secure Authentication & Data Layer
Implemented a PHP/MySQL backend for authenticated user management, historical log persistence, and secure session handling.
Key Architectural Decisions
Separation of ML Inference and CRUD Database
Decoupled the Flask inference microservice from the PHP/MySQL data tier to ensure model updates did not impact core user auth or vitals logging.
SwiftUI Dynamic Data Visualizations
Used native Swift Charts for fluid animations and instant feedback when users log new blood pressure or glucose readings.
Outcomes & Results
✦ Achieved 95% cardiovascular risk prediction accuracy across test validation datasets.
✦ Built a complete native iOS application with live charts and instant inference results.
✦ Integrated dual backends: secure user store (PHP/MySQL) and real-time AI serving (Flask).
Technical Reflection
“Building across iOS, Flask, and MySQL taught me the importance of API contract consistency and model interpretability in health-tech applications.”