Industrial biotechnology requires optimization strategies that can increase product formation while reducing process resource intensity. This study developed an artificial intelligence-assisted statistical optimization framework for recombinant enzyme production in a microbial host. The objective was to maximize enzyme yield and volumetric productivity while simultaneously improving a sustainability index based on energy and raw material efficiency. The empirical workflow combined fractional factorial screening, Central Composite Design, Response Surface Methodology, Artificial Neural Network modeling, and Genetic Algorithm optimization. Six operational parameters were first screened, and the statistically significant factors were then optimized using a response surface design. The same experimental dataset was used to compare a quadratic RSM model with a nonlinear ANN model. The ANN model showed stronger predictive performance than the RSM model, particularly for nonlinear regions of the design space. Genetic Algorithm optimization of the ANN response surface identified an operating condition that improved enzyme yield by 47% and reduced specific energy consumption by 22% relative to the baseline process. Bench-scale validation confirmed that the predicted optimum was experimentally reproducible. The study was limited to one microbial expression system and one recombinant enzyme product. The sustainability index was process-focused and did not include full cradle-to-gate life-cycle assessment. Pilot-scale validation remains necessary before industrial implementation. The originality of the study lies in embedding sustainability performance directly into an AI-assisted bioprocess optimization objective. By integrating DoE, RSM, ANN, and GA into a single sequential framework, the study demonstrates a practical route toward greener and more data-efficient biomanufacturing. The framework is intended to support sustainable biotechnology production under experimentally realistic constraints.
