Conventional hypoeutectic AlSi7Mg cast alloys (i.e. A356, A357 and variations) are widely used for structural castings in the automotive industry (e.g. engine blocks, cylinder heads) due to their low density and high specific strength that promote mass and consumption reduction, in view of a more sustainable mobility. An improved lightweighting can be obtained by combining the use of light metals like AlSi7Mg with topologically optimized designs and produced with innovative additive technologies such as laser-based powder bed fusion (PBF-LB). During their service life, automotive components are subjected to cyclic loading that can lead to fatigue damage and failure. While the fatigue behavior of conventional cast AlSi7Mg alloys is well established, the same is not true for these alloys processed by PBF-LB. The recent literature has proven that PBF-LB can be proficiently applied to AlSi7Mg alloys and most of the published works deal with their static behavior (i.e. hardness and tensile). Therefore, the present work aims at an extensive investigation of the low and high cycle fatigue behavior of the AlSi7Mg (A357) alloy produced by PBF-LB and subjected to tailored post-process heat treatments. Given the peculiar, extremely fine, and metastable microstructure induced by PBF-LB, heat treatments have been optimized to: i) induce an alloy strengthening by direct aging (T5 temper); ii) enhance the trade-off between strength and ductility by rapid solution treatment, quenching and aging (T6R temper). To assess the effect of tailored heat treatments on the fatigue behavior, S-N curve (stress vs number of cycles) and fatigue strength at 2x10^6 cycles were experimentally evaluated for the heat-treated alloy, while the as-built one was used as benchmark. The analysis of fracture surfaces led to the identification of killer defects that, despite the temper condition, in the majority of cases consisted in lacks of fusion. However, T6R treatment enhanced the overall fatigue behavior, outcome that was well correlated to the peculiar microstructural features.