The pursuit of fully virtualized product design workflows—encompassing initial design, manufacturing processes, performance prediction, and iterative redesign—has fueled rapid developments in multiscale and materials modeling. Concurrently, emerging machine learning (ML) techniques are being actively explored to accelerate these modeling workflows and reduce the time-to-solution for predicting complex material behaviors. Among these, deep material networks (DMNs) have emerged as a promising framework, offering fast and accurate nonlinear predictions across a broad class of material systems—including particle-reinforced composites, fiber-reinforced structures, and polycrystalline aggregates—even when trained solely on linear elastic data. Looking forward, the continued integration of mechanistic insights with data-driven learning will be essential for advancing the frontiers of multiscale modeling and enabling robust, generalizable simulations across engineering applications. This minisymposium aims to bring together researchers and practitioners at the forefront of computational mechanics, machine learning, and materials modeling. We welcome contributions that span fundamental developments, algorithmic innovations, and industrial case studies, with the goal of identifying key challenges and opportunities in deploying ML-enhanced multiscale methods in practical engineering workflows.