Techniques for processing input data for a detected user are described. Received image data is processed to identify an indicated user. Based on the user a machine learning model is implemented. The machine learning model is then used to process input data for a user input. An action is performed using the resulting output data.