Objective In this study, we aimed to maximize glutathione (GSH) production by a metabolically engineered Yarrowia lipolytica strain using a small-scale optimization approach. Results A three levels four factorial Box-Behnken Design was used to assess the effect of pH, inulin extract, yeast extract and ammonium sulfate concentrations on cell growth and to generate a mathematical model which predict optimal conditions to maximize biomass production and thus GSH titer. The obtained results revealed that only yeast and inulin extract concentrations significantly affect biomass production. Based on the generated model, a medium composed of 10 g/L of yeast extract and 10 g/L of inulin extract from Jerusalem artichoke was used to conduct batch cultures in 2 L bioreactor. After 48 h of culture, the biomass and the glutathione titer increased by 55% (5.8 g(DCW)/L) and 61% (1011.4 mg/L), respectively, as compared to non-optimized conditions. Conclusion From the obtained results, it could be observed that the model established from small scale culture (i.e. 2 mL) is able to predict performance at larger scale (i.e. 2 L bioreactor, two orders of magnitude scale-up). Moreover, the results highlight the ability of the optimized process to ensure high titer of glutathione using a low-cost carbon source.