The development of artificial intelligence (AI) educational technology presents new opportunities to improve the quality of learning, including in junior high schools with limited infrastructure. The main challenges for schools with limited infrastructure are limited accessibility to learning resources, a lack of teaching resources capable of providing personalized services, and limited adaptation of learning to individual student needs. Research focuses on the development of an AI-based learning platform capable of delivering a more personalized, efficient, and inclusive learning process. The planned platform aims to adapt content, learning methods, and learning pace to suit student profiles and needs. This study applies the Research and Development (R&D) method by adapting from Plomp which includes preliminary research, prototyping and assessment. Data collection techniques used interviews and observations in junior high schools in Lampung Province involving 28 students and 4 teachers as the main respondents. The results of this study using an AI-based algorithm, this system is designed to provide AI-based material recommendations to identify student weaknesses in teaching materials, synchronization features to update materials and questions when the platform is online, a student progress dashboard, and a collection of questions. Initial data shows the potential for increasing motivation and student engagement despite the limitations of school infrastructure conditions. This research provides added value to the development of inclusive, adaptive, and sustainable educational technology solutions, especially in the context of high schools with limited resources.