Evolution has created, selected and evolved large repertoires of proteins that operate in various biological systems. Nowadays biotechnological needs are coming up orders of magnitude faster than proteins naturally evolve. The emergence of de novo protein design accelerates the process of creating new proteins to explore a virtually infinite number of protein folds and sequences that can potentially be empowered with functionality. However, there remains a significant need for research in this field for installing and stabilizing structurally complex segments in designed proteins, and advance is needed in order to build and understand the structure–function relationship in the upcoming challenges in biological applications. As part of my thesis, I described two computational protocols to approach the design of proteins carrying irregular functional motifs from different perspectives; although the conceptual ideas varied, both methods eventually converged to construct the scaffolds for a function of choice while simultaneously optimizing protein stability. A particular field that will benefit from these computational approaches is the engineering of the epitope-focused immunogens for vaccine development. Previously, epitope immunogens have proved to be effective in eliciting potent neutralizing antibodies for an intractable pathogen: respiratory syncytial virus (RSV). The development of those immunogens relies on the transplantation of the epitope from its viral context into a heterologous scaffold for epitope stabilization. However, due to the lack of a generalized tool, the translation of complex molecular details into the design of effective immunogens is lagging far behind the available antibody–epitope information. Here, we borrowed the structural information of complex neutralization epitopes isolated from RSV and applied computational methods in order to engineer de novo proteins carrying irregular and discontinuous epitopes. Noticeably, we demonstrated that the structures of designed scaffolds closely resemble the design models, with sub-angstrom structural accuracy on the region of the epitopes. In vivo, we showed that cocktail formulations of three de novo designed immunogens presenting distinct epitopes induced RSV neutralizing antibodies in naïve nonhuman primates. Moreover, when pre-existing immunity is established, the designed immunogens are able to act as boosting immunogens to redirect the response onto the defined specificity. Also, we further functionalized the epitope-scaffolds as a BRET-based diagnostic biosensor to profile antibody responses in the serum, providing a valuable tool with which to quantify vaccination result with single epitope resolution. Beyond immunogen design, the use of our computational methods was applicable to assembling a protein topology bi-functionalized presenting two distinct binding motifs, which we applied as a non-natural inducer to control receptor-mediated signaling, and modulation of transgene expression in synthetic cells. Altogether, our advance in RSV-based studies demonstrates a substantial step for the utilization of epitope immunogens based on computationally designed proteins and may contribute to developing immunogens for other pathogens. The new methodological pipeline enables the design of a de novo protein with complex functional sites, which is a general blueprint for protein design to unravel new rules to create previously unimagined functional proteins.