The structural safety and performance of polycrystalline metal alloys is greatly affected by voids given their significant effect on the initiation and evolution of damage. Crystal plasticity theories in conjunction with finite element methods (i.e., CPFEM) are well-established computational protocols that can enable a fundamental understanding of the underlying behavior that relates the internal polycrystalline structure of metal alloys to its corresponding damage behavior and properties. However, despite their accuracy, CPFEM-based linkages are ill-suited to efficiently explore the extensive crystallographic orientation space because they are iterative in nature and need to solve numerically stiff, complex and non-linear constitutive equations which requires significant computational resources. Additionally, CPFEM-based linages are not capable of extracting useful knowledge from previously obtained results, as a new simulation is required for each void with a different local microstructure and/or loading conditions. Therefore, there is a critical need for computationally efficient linkages between the crystallographic orientation and the resultant damage behavior when a void is present. In this work, we circumvent this challenge by establishing a machine-learning-based linkage between the microstructure and the resulting damage performance. Specifically, we leverage machine learning model-building techniques to connect crystal orientations, represented with Generalized Spherical Harmonics, to the extreme values of the damage field in a single crystal when a spherical void is present. The resultant model is highly accurate and capable of predicting damage performance at a fraction of the cost. Furthermore, the insights and predictions obtained with the model showed great agreement with high-fidelity CPFEM simulation results. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525. SAND No. SAND2025-01952A