The use of current computer tools in both manufacturing and design stages breaks with the traditional conception of productive process, including successive stages of projection, representation, and manufacturing. Designs can be programmed as problems to be solved by using computational tools based on complex algorithms to optimize and produce more effective solutions. Additive manufacturing technologies enhance these possibilities by providing great geometric freedom to the materialization phase. This work presents a design methodology for the optimization of parts produced by additive manufacturing and explores the synergies between additive manufacturing, parametric design, and optimization processes to guide their integration into the proposed methodology. By using Grasshopper, a visual programming application, a continuous data flow for parts optimization is defined. Parametric design tools support the structural optimization of the general geometry, the infill, and the shell structure to obtain lightweight designs. Thus, the final shapes are obtained as a result of the optimization process which starts from basic geometries, not from an initial design. The infill does not correspond to pre-established patterns, and its elements are sized in a non-uniform manner throughout the piece to respond to different local loads. Mass customization and Fused Deposition Modeling (FDM) systems represent contexts of special potential for this methodology.