<p>This study focuses on addressing the supply chain network configuration problem by integrating two key dimensions namely responsiveness and resilience. To achieve this, a stagewise machine learning-based decision framework is proposed for identifying suitable transportation modes, selecting the best suppliers, selecting the best location to establish facilities and determining optimal flow of products. The main contribution of this study lies in developing an integrated stagewise framework that combines decision-making, machine learning, and optimization for resilient and responsive supply chain design, while also introducing a data-driven hybrid uncertainty modeling approach that links forecasting with multi-objective optimization. The initial phase involves evaluating potential transportation modes by scoring them against multiple criteria, employing the Fuzzy Best-Worst Method (FBWM) and Extremely Randomized Trees (ERT) methods. Following this, the second phase suggests a Multi-Objective Model (MOM) designed to configure and effective logistics system. To effectively manage the hybrid uncertainties inherent in the problem, a data-driven Fuzzy Robust Stochastic (FRS) optimization method is applied, combining fuzzy robust stochastic optimization with Seasonal Autoregressive Integrated Moving Average (SARIMA) forecasting techniques. Subsequently, a recently introduced solution approach named the Chebyshev Multi-Choice Goal Programming with Utility Function (CMCGP-UF) is used to find the optimal solution. Results from the first stage highlight that Van and Middle size truck stand out as the most appropriate transportation modes. Also, findings reveal that increasing demand leads to a rise in all objective function values. The analysis also shows that higher value for the service level has led to increasing the total cost.</p>