Nonlinear interface phenomena govern the behaviour of many biomechanical systems, from soft robots in viscous flow to braided endovascular implants in the brain vasculature. Yet, they are often tamed numerically by overly stiff coupling or simplified contact models. This contribution connects two such problems at very different scales (see Fig. 1): soft robotic swimmers (SRS) and braided endovascular devices (BED) for cerebral aneurysm treatment [1-3]. In the first part, we use a 3D FSI framework to model nature-inspired magnetic SRS, where hard-magnetic elastomer bodies are driven by rotating fields and coupled to viscous flow. Numerical simulations show how the balance between remanent magnetization, flexural rigidity and fluid viscosity selects propulsion regimes, efficiency and emergent collective patterns in swimmers and swarms. In the second part, we simulate a specific BED (the Contour device) as parametrically generated wire architectures discretized with geometrically nonlinear beam elements and deployed into patient-specific, deformable aneurysm sacs. Explicit frictional contact is resolved at wire-wire crossings and along the device-wall interface, enabling in silico crimping, catheter delivery, and gradual release. Across parameter sweeps, friction emerges, after material stiffness, as a key control knob for expansion kinetics, final apposition and post-release positional stability, in sharp contrast to the over-constrained response obtained with kinematic ties. Finally, using these two case studies, we claim that a rigorous modelling of nonlinear interfaces, i.e., FSI for SRS and frictional beam-beam/beam-surface contact for BED, is essential for high-fidelity computational biomechanics.