We propose a research trajectory in the field of deep learning applied to music production systems such as mixing, mastering, sound design and sound synthesis.