Multiscale modeling and simulation of blood flow in the human circulatory system can give an interpretation about the interaction between complex processes that occur at different scales determined by the flow features in the large and medium size arteries at a macroscale level (diameter of 0.1 cm or larger), in the smaller arteries and arterioles at mesoscale (mean diameter of 50 μm) and in the capillaries of the microvasculature (mean diameter of 800 μm). For instance, at the macroscale, in addition to the complexity of the vascular geometry, blood may be considered as a Newtonian or a non-Newtonian fluid, depending on the size of the vessel, which interacts with the vessel wall, resulting in pressure waves that deform under the action of blood pressure. Appropriate 3D fluid-structure interaction (FSI) models need to be considered to locally represent these phenomena, while reduced 1D (distributed parameter) and 0D (lumped parameter) approximations are used to account for the remaining parts of the systemic circulation (geometric multiscale). Several challenges arise when coupling continuum macroscale with atomistic meso- and microscale models. The most relevant are the design of efficient and robust interface conditions and the accuracy of the results obtained from the interaction between the local small-scale resolutions with the global ones. The main goal of this talk is to present an overview of some approaches in multiscale modeling and simulations of blood flow problems. Moreover, we will discuss the issue of integrating machine learning and multiscale modeling in biomedical applications, with special focus on blood flow problems.