The transition to battery electric vehicles requires electrification of power systems. The reliability and efficiency of electrical conversion systems are often compromised due to inadequate thermal management systems, leading to electronic failures and performance degradation. Addressing this challenge necessitates the optimization of thermal systems and the development of innovative vehicle architectures. The proposed study utilizes a novel design framework for topology and sizing optimization of electric vehicle energy systems using a graph-based modeling approach. To address the increasing complexity of electric vehicle systems, the model incorporates multi-physics modeling, integrating the interaction between the thermal, electrical, and mechanical domains. Additionally, the framework uses energy conservation laws to capture physical system dynamics, making optimization more structured and helping analyze design trade-offs effectively. To demonstrate its effectiveness, an electric bus powertrain design is investigated as a case study, with the goal of optimizing thermal and electrical component sizes and energy flow, as well as discrete choices in the topology of the system. The electric bus model incorporates experimentally validated empirical data obtained through advanced measurement techniques, ensuring credible optimization and realistic system behavior. This case study demonstrates how design optimization can influence the efficiency of thermal management systems.