HandPose is a machine-learning model from ml5.js that detects hand positions and landmarks in real-time using a webcam. Built on TensorFlow.js, it identifies 21 key points on a hand, including the fingertips and joints. Additionally, ml5.js enables the creation and training of custom machine learning models directly in the browser using its Neural Network capabilities. In this experiment, we trained the model to recognize specific hand gestures, which then dynamically update the variable font axes and interpolations.