How can structured knowledge networks enhance the accuracy and relevance of LLM-generated responses within human learning systems? This study investigates integrating graph-based knowledge representations with LLMs to support personalized and adaptive learning experiences. We model curated educational content as a structured knowledge graph. Then, leveraging graph-based Retrieval Augmented Generation (GraphRAG) techniques, we explore how the topology of generated knowledge graphs improves LLM response quality and the generation of personalized learning segments. Our approach enables dynamic alignment between student queries and extracted subgraphs of domain-relevant content, which improves the contextual grounding of LLM outputs. We evaluate the method within the linear algebra content domain, demonstrating improvements in response accuracy, usefulness, and completeness compared to baseline retrieval methods. Beyond personalized education, this study positions graph-augmented LLM frameworks as a generalizable mechanism for navigating structured knowledge networks in human-AI collaborative systems. This research advances the discussion on optimizing learning pathways by demonstrating the effectiveness of knowledge graphs applied to LLMs for generating personalized learning segments.