ABSTRACT:  Understanding boundary layer flows over rough surfaces is important in engineering problems. Processes such as bio-fouling on marine vessels, or erosion on gas turbine blades form irregular roughness topographies, which increases the friction drag and decreases the efficiency of the engineering systems. Roughness-resolved simulations, such as direct numerical simulations (DNS), play a key role in providing physical insights into the roughness effects on turbulent flows. They are challenging due to the large range of space and time scales that need to be resolved. In my doctoral research, I developed a code on parallel platforms and conducted DNS to investigate roughness effects on transitional and turbulent boundary layers. The first part of my work focused on turbulent flows over rough surfaces with random geometries. I did in-depth analysis to understand the impact of random rough surfaces on the pressure and wall-shear stress fluctuations. The second part of my work focused on the influence of different roughness geometries and distributions on laminar-to-turbulence transition. I utilized DNS in conjunction with global linear stability tools for the analysis and identified the joint effects of roughness characteristics and flow conditions on the instability mechanisms and transition process. The outcome of this study provided fundamental insights into how to design and optimize roughness geometries in engineering applications.