This is a EU-funded initiative to combat global dengue epidemic. The early detection of dengue patients at risk of developing severe outcomes is crucial for managing healthcare resources effectively, mainly to prevent hospitalization. Despite the identification of specific biomarkers that are indicative of acute inflammation and hepatic involvement, there remains a significant need to develop reliable predictive tools for early intervention in dengue management. Lack of uniform and structured protocol, a handful of analyte assessments, inappropriate study design, and application of computational tools further posed a severe challenge to pinpointing the early predictive biomarkers for dengue severity. The prevalence of different Dengue virus serotypes (DENV-1-4) can vary between countries, affecting the severity of outbreaks and the population's immunity. Moreover, a recent meta-analysis of the dengue severity biomarkers indicates the heterogeneity in the biomarker prediction. Therefore, in the context of diseases like dengue that have diverse epidemiological patterns across different regions, a single center or regional study can introduce several gaps and limit the ability to generalize results to broader populations. Addressing these gaps, we aim to conduct multi-site studies that encompass a range of geographical locations, socio-economic contexts, and public health environments to develop AI tools to enhance the generalizability, accuracy, and fairness of AI applications. This approach enhances the robustness, relevance, and applicability of research findings to a broader spectrum of settings and populations.