Enzymes are increasingly combined into multienzyme systems for cost and productivity benefits. Further advantages can be gained through the use of immobilized enzymes, allowing continuous biotransformations in flow. However, the optimization of such multienzyme systems is challenging, particularly where immobilized enzymes are used. Here, we meet this challenge using both mechanistic and empirical modeling to optimize a reaction involving a reductive aminase and a glucose dehydrogenase for continuous biocatalytic reductive amination in flow. Crucially, the construction of the mechanistic model was achieved quickly, with only a few important parameters determined experimentally, and ensemble modeling used to facilitate the use of estimates or literature values. Upon reaching the limits of the mechanistic model's capabilities, we show that solution space can be further explored using a definitive screening design to generate an empirical model of the reaction, using the mechanistic model's prediction as a starting point. We demonstrate the use of this empirical model to design optimal processes for either high productivity or to minimize necessary cofactor and cosubstrate concentrations. Our results demonstrate that the synergistic use of both mechanistic and empirical modeling offers a route for rapid optimization of multienzyme systems of immobilized enzymes in flow with minimal experimental effort.