Scalable adaptive assessment framework for collaborative eLearning: a service-oriented AI approach

Abstract Traditional eLearning assessments frequently rely on static, non-adaptive methods that provide limited personalized feedback and face significant scalability constraints. The Service-Oriented Framework for Intelligent Assessment (SOFIA) integrates Service-Oriented Architecture (SOA) principles with AI-driven adaptive feedback mechanisms to deliver real-time, personalized assessments at scale. Its modular, microservices-based architecture enables seamless integration with Learning Management Systems (LMS) and supports a wide range of educational contexts. The framework processes real-time learner interaction data using priority-based task optimization algorithms and generates iterative feedback informed by performance metrics. An experimental study involving 250 participants in an online Python programming course demonstrated its effectiveness. SOFIA achieved 92% accuracy in identifying learning gaps, outperforming static assessments (65%) and rule-based adaptive systems (78%). Learner engagement increased by 40%, and post-test retention scores improved to 85% compared with 65% in the control group. Scalability tests confirmed the framework’s capacity to efficiently support more than 500 concurrent users. SOFIA addresses key limitations of conventional eLearning assessments by enhancing engagement, performance, and retention. It provides a scalable and adaptive solution that integrates seamlessly into existing LMS platforms, equipping educators with automated personalized assessment and feedback tools. While the study focused on Python programming, future research should examine its applicability across broader disciplines and pedagogical models. This work establishes a strong foundation for next-generation intelligent educational technologies by combining adaptive assessment with service-oriented design principles.