We present the design and performance evaluation of a novel multisorbent process for CO2/N2 separation based on vacuum-swing adsorption (VSA). We study two process configurations: (i) layered-bed processes, wherein two distinct adsorbent materials are arranged in sequential layers within the adsorption bed, and (ii) mixed-bed processes, wherein two distinct adsorbent materials are homogeneously mixed within the adsorption bed. We develop, validate, and deploy a high-fidelity dynamic adsorption column model for the multisorbent process configurations and apply Bayesian optimization to design processes that achieve maximum separation effectiveness in terms of CO2 purity and recovery with an application to postcombustion carbon capture (PCC) on a coal-fired power plant. We find that the multisorbent process configurations achieve improved CO2/N2 separation effectiveness compared to benchmark classical single-adsorbent processes, increasing the CO2 recovery by up to 5% while achieving high CO2 purity. When operating in compliance with widely adopted performance targets for PCC (PuCO2 ≥ 95%, ReCO2 ≥ 90%), we find that the multisorbent process configurations reduce the energy usage of the separation by approximately 35%. We use the modeling framework to analyze the subcolumn scale adsorption dynamics and identify that the observed improvements in performance are associated with the positioning of the CO2 adsorption front under optimized operating conditions, leading to favorable dynamic interactions with the operation of the VSA process cycle.