As an extension of my individual final year university project, I developed a real-time gym exercise detection app that has a 99.9% accuracy using the time series classification model I developed. The app records accelerometer and gyroscope sensors in an Android smartphone at a frequency of 10 Hz and records 15 seconds of data. To create the app and access the phone's sensors, I utilised Android Studio Java. In addition, I set up a Python Flask server that ran my trained machine learning model, allowing for real-time predictions within the app by sending API calls to the Flask server.