Master thesis is focused on the investigation of the initial fatigue life stages of austenitic stainless steels with the accent on characterising microstructural evolution, such as cyclic plasticity and crystallographic character. The initial stage of cyclic loading is a critical moment that defines the subsequent cyclic stress-strain response and residual lifetime. Therefore, a detailed understanding of the microstructure-mechanical properties relationship is crucial for the design of new materials or engineering structures. For this reason, an interrupted fatigue test on AISI 304L grade steel was performed under a strain-controlled regime with a symmetric loading cycle, to capture the development of cyclic plastic deformation, crystallography, phase composition and stress state during the first fatigue loading cycles by digital image correlation via scanning electron microscopy (SEM-DIC), electron backscatter diffraction (EBSD), high-resolution (HR)-EBSD and feritscope. To ensure successful test execution, it was necessary to design and test the specimen geometry and preparation methodology, as well as the fatigue test procedure. After the fatigue experiment, the chemical composition and topography were analysed by energy-dispersive spectrometry (EDS) and in situ atomic force microscopy (AFM). Fatigue behaviour consisted of initial cyclic hardening, followed by saturation cyclic softening, which progressively weakened until it reverted into secondary cyclic hardening. The latter fatigue stage was accompanied by the evolution of negative mean stress. SEM-DIC revealed the gradual activation of slip bands, which correlated well with the initial cyclic hardening stage and loading conditions. Local misorientations were revealed, originating from grain slip activity and the influence of surrounding grains. Strain-induced martensite nucleation during secondary cyclic hardening was confirmed by feritscope measurements. Revealed chemical heterogeneity led to SFE variability and promotion of martensitic transformation in the low-SFE areas. All the analytical techniques were found to be in good agreement and provided deep, mutually complementary insights.