In this study, a multi-scale approach was used to improve the production titers of two classes of heterologous natural products in Escherichia coli: the polyketide 6-deoxyerythronolide B (6-dEB; complex precursor to the antibiotic compound erythromycin) and the diterpenoid taxadiene (complex precursor to paclitaxel). These studies included model-driven analysis of E. coli metabolism to identify knockout candidates that are predicted to improve production, native and heterologous gene over-expression to improve the uptake and utilization efficiency of exogeniously fed substrates, the development of a novel computational algorithm utilizing elementary mode analysis and a genetic algorithm and its application towards improving biosynthesis, and process and cellular engineering strategies to improve both volumetric and specific productivities.