Large-scale structure surveys allow us to constrain cosmology through the growth of structure and the expansion history. Together with Cosmic Microwave Background (CMB) measurements from the early Universe, the standard CDM with general relativity has become a concordance model that explains the origin and evolution of our Universe for over thirteen billion years. However, recent analyses reveal hints of tension between the early and late time Universe observables. The tension could point to newphysics beyond CDM, but could also be due to unknown systematic biases. Improving the precision and accuracy of cosmological measurements could be the key to a revolutionary new discovery about gravity on cosmological scales or about the nature dark energy. The challenges of cosmology in this era is to broadly explore and control possible systematics that could limit science from on going Stage III and future Stage IV surveys. In this thesis, we investigate astrophysical systematics of intrinsic alignment (IA) and the baryonic physics that would contaminate the weak lensing observables. Intrinsic alignment of galaxies arise under the effect of their local gravitational field. This effect mimics the shear correlation and introduces a systematic bias in the measured weak lensing signal.We carried out IA analyses using 8000 redMaPPer clusters. There are two types of alignment withinthe one-halo scale: the alignment of the central galaxy with respect to the host halo shape and the radial alignment of satellite galaxies toward the halo center. For the central galaxy alignment, the mean misalignment angle between the central and cluster major axes is measured to be 35 (random alignments give 45 ). The satellite galaxy alignment is a more subtle effect. We concluded that no net radial alignment signal is detected across the entire sample based on the re-Gaussianization shapes (the most conservative measurements). We also studied the dependence of IA signal on a total of 17 cluster and galaxy related properties in a concordant framework to properly account for parameterdegeneracies. With several predictors identified for central and satellite alignments inside galaxy clusters, our results suggest that small-scale IA is a complicated phenomenon potentially involving multiple relevant physical processes during galaxy and cluster formation and evolution history.The modification of the matter distribution due to baryonic physics is a non-negligible source of uncertainty for precision cosmology. Using various sets of hydrodynamical simulations, we investigated the effect of baryons on the matter power spectrum as functions of wavenumber frequency and redshift. We developed methodology to model the effect of baryons on weak lensing cosmic shearobservables through the principal component analysis (PCA). We constructed mock cosmic shear observables with contamination from different hydrodynamical scenarios, and then performed PCA on the contaminated mock observable vectors. The resulting principal component (PC) modes are then form a set of efficient bases to span uncertainties of baryons in the observable data vector space.Overall, our results suggest that the effects of baryonic physics on cosmic shear power spectra can beefficiently captured and mitigated using a few linear combinations of PC modes. Small-scale information in galaxy surveys has substantial statistical power to improve cosmological constraints, but conventional cosmological analyses discard this information to avoid biased inferenceon parameters due to lack of good astrophysical models for it. The continued developments on astrophysical modeling techniques in advance with better statistical quality of data will be essential in the new era of precision cosmology.