Generating Faculty Specific Course Materials With Neo4j and Claude
summary
Project converts 20 years of course transcripts into a searchable Neo4j knowledge graph that links faculty-verified topics to transcript chunks, material nodes, and generated learning objects; it uses Hugging Face instruct embeddings, LangChain, Python, Neo4j vector indexes and AWS Bedrock + Claude for LLM runs. LLM outputs are merged across runs, distilled to 25–30 vetted topics per course, turned into flashcards (publishable to Canvas) and tracked with similarity scores and study-session logs. System is live and faculty-approved outputs are used and refined.
AI-written summary · about
chapters (12)
- 0:00Introduction and team background
- 1:56UD Study Aid project overview
- 4:27Building the knowledge graph structure
- 8:05Faculty review and data verification
- 10:30Prompt engineering and testing
- 13:21Faculty interface and approval process
- 14:48Code walkthrough and graph implementation
- 17:37Learning objects creation and student interface
- 21:35Generating learning objects from topics
- 23:02Future directions and new data inputs
- 24:52Tracking study sessions and student progress
- 25:57Q&A and university collaboration
related talks
44mGive Your LLMs a Left Brain
14mLLMs and Knowledge Graphs: A Case Study in Blue(y)
36mGraphRAG Explained 🔥 Building Smarter LLMs
10mTalk to Your Graph: A Practical Guide to Building a Dual-LLM Q&A System for a Fashion Graph
28mGrounded Intelligence Boosting LLMs with Real-Time Graph Context
01h 17mGraph Database and Retrieval Augmented Generation
22mBuilding Contextual Knowledge Graphs to Adapt, Evaluate, and Leverage LLMs
52mGive Your LLMs a Left Brain