I'm Roaa
Khaled.

Open-loop insulin delivery system integrating ESP32 hardware, a Dense neural network (96.53% accuracy, 17.3 KB on TFLite) forecasting blood glucose 30–60 min ahead, and a Flutter companion app with role-based dashboards for patients, doctors, and guardians.

Flutter mobile app for an autonomous residential gate system with face recognition, anti-spoofing, and gesture detection on ESP32-CAM. Features remote gate control, live feed, two-way intercom, guest OTP generation (6-digit, 15-min expiry), and rotating resident PINs via MQTT.

Full ML pipeline including image acquisition, preprocessing (Gaussian blur, gamma correction), HSV segmentation with morphological operations, and feature extraction. Trained Logistic Regression and Random Forest classifiers, evaluated with confusion matrix and ROC curve.

Classification model for predicting credit card approval using feature selection with Genetic Algorithms, trained with multiple ML classifiers and evaluated for performance.

Team-built music streaming platform featuring song recognition, personalized playlists, and an admin management panel. Contributed frontend and backend development.

Course recommendation system with a chatbot interface built using an Agile workflow. Integrates JavaScript frontend with PHP backend to deliver personalized learning paths.