Machine learning is becoming more and more popular for screening for genetic targets, as it can decrease the number of combinations to be studied. Here we combined in a two-step approach mechanistic and machine learning models to increase tryptophan production in yeast, as a proof of concept. In the first step we used genome-scale modeling to pinpoint the genes to focus on, and in the second step we used machine learning for predicting the most promising promoters for said genes, using fluorescent data from a tryptophan biosensor. I was in charge of the genome-scale modeling step of the approach, which was the same as the method introduced in a previous publication. Running this approach gave us a strain that increased tryptophan titer and productivity by up to 74% and 43%, respectively, so it presents as a promising tool for metabolic engineering.