While conventional manufacturing and design methods of composite laminates maintain both fiber angle and thickness constant within a ply, tailored fiber placement (TFP) process allows producing laminates by steering fibers curvilinearly, generating structures with a variable-stiffness characteristic. This offers more flexibility to tailor the mechanical properties of laminated structures over conventional constant-stiffness ones. This study presents a methodology to optimize an anisotropic composite structure, comprising in performing cross-section optimization of a topologically-optimized structure through an evolutionary optimization using a genetic algorithm (GA). The optimization formulation is built up accounting for manufacturing characteristics of the TFP process by imposing constraints to generate fiber patterns feasible to be manufactured. The proposed approach locally optimizes both fiber angle and intrinsic thickness build up simultaneously. The structure with its topology and cross-section optimized has a specific stiffness 330% higher than the quasi-isotropic stacking sequence. The cross-section optimization enhances the specific structural stiffness in 22% compared to the topologically-optimized structure.