Chiral mechanical metamaterials are architected structures with broken centrosymmetry, inducing coupling between axial strain and rotational deformation. Their compression–twist response arises from linear and nonlinear mechanisms that are highly sensitive to unit cell topology and geometric parameterization. Conventional chiral designs typically exhibit fixed deformation modes and limited tunability, constraining performance and requiring extensive computational exploration. Consequently, a systematic, mechanics-based design strategy is needed to identify the geometric factors governing chirality amplification and to efficiently predict twist-dominated behavior across diverse architectures This work introduces a programmable chiral unit cell capable of selectively activating multiple deformation mechanisms, enabling customizable chirality and a substantially broader operational range than standard designs. A computational mechanics framework is developed to model programmable chiral systems, allowing transitions among bending-dominated, rotation-dominated, and instability-assisted responses without altering topology. Finite element simulations in COMSOL Multiphysics include mesh convergence verification and parameter sensitivity analysis, generating a large dataset spanning wide geometric variations to support data-driven modeling. Key performance metrics, including effective elastic modulus, relative density, and twist-to-strain ratio are evaluated to establish predictive insight into mechanical behavior. Correlation and clustering analyses identify the geometric factors responsible for chirality amplification, revealing fundamental structural elements that enable unit cells with twist-to-strain ratios approaching or exceeding 10. Computational predictions are validated through additively manufactured specimens, showing close agreement between experimental and simulation results. Finally, a physics-informed neural-network surrogate combined with inverse design optimization enables rapid exploration of the high-dimensional design space and efficient discovery of unit-cell geometries meeting prescribed mechanical criteria, building on recent advances in physics-guided machine learning for inverse design of cellular materials.