Academy Research Papers

Learning That Remembers You: An AI-Native Ecosystem for Adaptive, Memory-Aware, and Multimodal Education at Scale

Dhanikesh Karunanithi · Sudar / ALP Project · 2026

Traditional learning management systems (LMSs) deliver static, one-size-fits-all content with no longitudinal learner model and no adaptive tutoring. Intelligent tutoring systems (ITS) that do adapt remain either narrow-domain research prototypes or products disconnected from the course-hosting infrastructure most organisations already use. We present Sudar, an AI-native learning system with three main contributions: (1) a full open-source reference platform that unifies authoring, delivery, and intelligence around a persistent Digital Learner Twin, adaptive sequencing, a complete multimodal delivery stack (text with read-along TTS, video, audio podcast, mindmap, flashcards, and SCORM), and an AI tutor with longitudinal cross-session memory and proactive nudges, where learners and organisations can govern how often LLM inference updates the stored learner model and digest summaries (data minimisation); (2) the Adaptive Learning Layer (ALP), an architecture by which these capabilities can be deployed as standalone plugins on top of existing LMSs; and (3) a demonstrated radical cost efficiency enabled by open-weight inference models and zero-cost TTS, making world-class AI-native learning economically viable at a per-learner infrastructure cost of less than $0.02 per month. The reference implementation is open source under the Apache License, Version 2.0, evidence-informed by the learning sciences.

@misc{sudar2026,
  author = {Karunanithi, Dhanikesh},
  title  = {Sudar: An {AI}-Native Learning {OS} and Adaptive Learning Layer},
  year   = {2026},
  url    = {https://github.com/Dhanikesh-Karunanithi/Sudar},
  note   = {Reference implementation, Apache-2.0 licence}
}