The muse of this project is Glass, the Brittle Beauty. Computational tools are adapted and assembled into an optimization workflow to inform the structural design of cast glass. The objective function is the weighted-sum of three components: 1) area ratio, and the p-norm of both 2) stress and 3) displacement. The Drucker-Prager stress criteria for brittle materials is used to calculate the stress component. There is a soft constraint on annealing time and even thickness embedded in the material distribution definition, which also has a considerable reduction in the number of parameters and post-processing effort needed when compared to the SIMP methodology. This material distribution definition relies on double slicing Gaussian Random Fields with a transformed sigmoid curve. The optimization algorithm is written in python. It makes use of pytorch’s automatic differentiation functionality to calculate the gradients, and its Stochastic Gradient Descent (SGD) and Adam optimizers to update the material distribution parameters. It culminates in the design of an indoor pedestrian bridge with significantly lower tensile stresses than comparable projects have achieved, with appreciably lower computing power used.