Architected metallic metamaterials fabricated by additive manufacturing are called to expand infinitely the variety of available properties observed in bulk alloys. However, the high surface-to-volume ratio of the architected metamaterials due to their intricate geometries and the surface inherited of the AM process is translated to a complex fatigue behaviour when compared with bulk conventional alloys. This is a serious concern in the use of this new class of architected AM materials in technological applications. In this work, this problematic is tackled by a systematic multiscale study of the metamaterial design – microstructure and defects – fatigue properties’ interconnection. Commercial aluminium and titanium alloys processed by selective laser melting are used as base material. By means of combined fatigue experimentation, computational modelling and machine learning approaches, the effect of processing conditions and design geometry on microstructural defects and surface quality is rationalised and connected with the fatigue life of metamaterials. (Proyecto PID2020-116440RA-I00 financiado por MCIN/AEI/10.13039/501100011033)