-Enhanced functionality of a wearable application for detecting eye-rubbing by developing a custom Transformer model used for binary classification, enabling accurate identification of instances critical for patients diagnosed with Keratoconus (a progressive eye disease).-Developed two applications, including a data collection tool utilized to gather two million data points and a deployment platform for the optimized model, resulting in a robust and accurate solution addressing previous limitations.-Acquired and applied expertise in a diverse range of AI models including RNNs, LSTMs and Transformers, employing advanced techniques such as fine-tuning and data cleaning to optimize performance.-Executed the deployment of AI models onto wearable platforms, including WatchOS and WearOS, using TFLite and Core ML, ensuring seamless integration and optimal performance.