AI-Powered Real-Time Food Prediction App

I developed a real-time mobile application in Swift that leverages machine learning to predict food types from audio input, demonstrating strong technical skills in mobile development and AI system integration. Working with a substantial dataset of 11,000 samples across 20 categories, I built and optimized machine learning models using Python and CoreML, employing Random Forest algorithms to achieve 84% training accuracy validated through cross-fold validation techniques. This project required rapid prototyping and iterative development, as I systematically modified hyperparameters, including window sample size and overlap factor,s to enhance model performance. The work showcased my ability to bridge research methodologies with practical product implementation, translating complex machine learning concepts into a functional user-facing application that processes real-time data streams.