Domanda di colloquio di Infineon Technologies

How would you optimize a deep learning model to improve its performance without overfitting?