<![CDATA[The rapid growth of large-scale hybrid learning environments has increased the need for data-driven approaches to identify and support at-risk students before disengagement or dropout occurs. Many institutions struggle to respond proactively due to the absence of predictive mechanisms that translate real-time learning data into actionable interventions. This study investigates the use of predictive analytics and artificial intelligence (AI) for early identification and intervention among at-risk students within a large-scale hybrid university program. The research aims to evaluate how machine learning models can detect behavioral and academic risk patterns and how these predictions can inform timely academic support strategies. A quantitative predictive research design was employed using secondary data from the university’s learning management system (LMS) and student information records. Data from 5,000 hybrid learners were analyzed using regression-based predictive models and supervised machine learning algorithms, including random forest and logistic regression, to determine key predictors of risk. Validation was conducted through cross-validation and accuracy metrics. The results revealed that engagement frequency, assessment completion rate, and login regularity were the strongest predictors of student risk, with predictive accuracy reaching 89%. Early interventions informed by predictive insights such as personalized feedback and AI-assisted tutoring led to a 23% reduction in course withdrawal rates. The study concludes that predictive analytics and AI can significantly enhance institutional capacity for proactive intervention in hybrid education. The integration of automated early-warning systems represents a transformative approach to promoting equity, retention, and personalized learning support at scale]]>
