🧪 🔬 Presented work

Paper 1. «Autonomous Syllabus Generation via LangGraph ReAct and Hierarchical Agentic RAG»
Authors: Gulshat Amirkhanova, Bauyrzhan Amirkhanov, Alikhan Amirkhanov, Ramilya Aubakirova
Summary: An agent-based system for automatically generating syllabi and course structures from unstructured documents based on LLM and hierarchical search with augmented generation (Agentic RAG). The LangGraph ReAct orchestrator operates without user-defined targets, using a multi-tool extraction strategy. For the Occupational Safety and Health (OSH) project, the system organized the material into four thematic modules and 16 lessons.

Paper 2. «Adaptive Prompting: Using Large Language Models to Generate Real-Time, Personalized Scaffolding for STEM Students»
Authors: Bauyrzhan Amirkhanov, Azim Aidynuly, Gulshat Amirkhanova, Alikhan Amirkhanov, Dinara Zhaisanova
Summary: An adaptive prompting system that links cognitive diagnostic models to a student’s interaction database and implements pedagogical constraints into the LLM prompt, forcing the model to provide hints and leading questions instead of ready-made answers. Tested on 120 students in a higher mathematics course, users of the adaptive system showed higher results on the post-test.

Paper 3. «Automation of Generation of SCORM-Standard E-Learning Courses Based on Large Language Models and Retrieval-Augmented Generation Technology»
Summary: A system for fully automated generation of SCORM 1.2-standard training courses using LLM and RAG. The focus is on simplifying the work of methodologists by automating the labor-intensive process of course creation.