Additive manufacturing (AM) enables the direct fabrication of complex geometries from digital models, while providing precise control over processing parameters to tailor the mechanical performance of polymer-based materials. Fused deposition modelling (FDM) is among the most widely used AM techniques due to its accessibility, flexibility, and ability to produce intricate structures. However, the mechanical behaviour of FDM-printed components is highly sensitive to process parameters, resulting in anisotropy and variability in mechanical properties that challenge reliable engineering design. This study investigates the process-structure-property relationships of Onyx-FR, a recently developed flame-retardant polyamide 6 reinforced with short carbon fibers by Markforged, for which process-specific mechanical data are scarce. Specimens were fabricated under systematically varied printing conditions, including four raster angles, two build orientations, four layer heights, and two extrusion temperatures, yielding 40 unique configurations. Quasi-static tensile tests were conducted following DIN EN ISO 527, and the resulting stress-strain data were used to calibrate a Ramberg-Osgood (RO) constitutive model. A MATLAB-based optimization procedure was developed to achieve the best-fit curves and extract the RO parameters, establishing a generalized relationship between the strain-hardening parameters and layer height. Fracture surfaces were analysed to identify the dominant failure mechanisms and their correlation with printing-induced microstructural features. Results show that raster angle is the primary factor controlling stiffness, strength, and ductility, while layer height and extrusion temperature exhibit secondary effects. Pronounced anisotropic behaviour reflects layer-wise deposition, interfacial bonding, and void formation. The RO-model accurately predicts material response across different printing conditions. Fracture analysis highlights the critical role of interlayer adhesion and fiber-matrix-interactions in governing failure. Overall, this work establishes a combined experimental and predictive modelling framework for FDM-printed Onyx-FR composites, providing data-driven insights into their mechanical performance. The findings support performance-oriented design of AM-components and broader application of advanced polymer composites in engineering systems.