Data Sheet 1: Multi-Agent RAG System for SCORM Course Generation

Gulshat Amirkhanova's research document describes a multi-agent pipeline for generating SCORM 1.2-compliant e-learning courses from enterprise documents. The system uses large language models and retrieval-augmented generation (RAG) to automate course design and content creation. The document, last updated on 2026-06-01, is 679.8 KB in size and shared under a CC-BY-4.0 license.

Use Cases

• Developing automated course generation systems based on the described multi-agent RAG pipeline.
• Benchmarking AI-driven instructional design systems against the semantic document ingestion and structure-aware chunking methods outlined.
• Implementing standards-compliant SCORM packaging for learning management systems based on the described pipeline.
• Researching methods to reduce hallucinations in AI-generated educational content by ensuring traceability to source materials.

Strengths

• The system was evaluated using real-world occupational safety documents, suggesting practical application.
• The document provides a detailed, four-stage technical description of the pipeline, including semantic ingestion, autonomous architect agent, parallel content generation, and SCORM packaging.
• The dataset is shared under a permissive CC-BY-4.0 license.

Limitations

• The dataset is a 679.8 KB DOCX file, which is a very small document describing a system rather than a substantial data corpus.
• Row count and column-level documentation are absent; the actual data content and structure must be inferred from the description.
• The description is detailed but does not provide access to the underlying code, models, or generated course data.

Provenance

Source

Gulshat Amirkhanova via figshare.

Collection Method

Likely represents research output documenting a developed system.

Freshness

Last updated 2026-06-01 04:18:55; freshness should be verified.