Generative machine learning methods that use neural networks to parameterize large-scale and near genome-scale kinetic models have delivered significant efficiency gains in model construction, paving the way toward high-throughput dynamic metabolism studies for biomedical and biotechnological applications. Nevertheless, challenges remain in interpreting the outputs of generative neural networks and developing strategies to quickly adapt these networks to different organisms and physiological contexts without restarting the modeling process from scratch. Here, we present a systematic framework for repurposing generative neural networks trained on one physiological context to build large-scale kinetic models tailored for another context, thereby offering a new avenue for efficiently constructing models with targeted desired properties suitable for various physiological scenarios. We showcase the effectiveness of this method through three case studies: (i) adjusting the response speed of cellular metabolism in aerobic E. coli cultures, (ii) improving interpretability by identifying key enzymatic steps that limit the speed of metabolic responses, and (iii) adapting our neural network to capture the distinct dynamic behavior of anaerobic E. coli. Given the growing adoption of generative neural networks in biological systems modeling, our approach has the potential to advance personalized medicine and accelerate the high-throughput design of cell factories by streamlining model construction across diverse living organisms.