Agentic GraphRAG Recommender System
Agentic GraphRAG Recommender System
Overview
Published at the AAAI 2025 Spring Symposium (Human-Compatible AI) and supported by the Robert Mayer Undergraduate Research Fellowship, this project develops an autonomous multi-agent advising architecture designed to solve complex combinatorial university degree planning challenges. By fusing Knowledge Graph representations (Neo4j) with dense semantic vector retrieval (Qdrant) and vLLM continuous batching, the system eliminates prerequisite hallucinations and slashes advising turnaround times.
flowchart TD
Student[Student Preferences & Transcript] --> Agent[Orchestration Agent\nLangChain / Tool Calling]
Agent <-->|Graph Traversal\nPrerequisites & Constraints| Neo4j[(Neo4j Knowledge Graph)]
Agent <-->|Dense Semantic Search\nCourse Descriptions| Qdrant[(Qdrant Vector DB)]
Agent <-->|Fast High-Throughput Inference\nContinuous Batching| vLLM[vLLM Inference Engine]
Agent --> HITL[Human-in-the-Loop (HITL)\nAdvisor Verification Portal]
HITL --> Plan[Optimized Graduation Plan]
System Highlights & Innovations
1. Hybrid GraphRAG Architecture
- Modeled complex university curriculum catalogs, prerequisite DAGs, core requirements, and major/minor electives into Neo4j.
- Leveraged Qdrant dense vector indexing to semantically align student interests, career trajectories, and course descriptions.
- Hybrid retrieval prevents standard LLM hallucinations regarding course prerequisites and degree credit requirements.
2. High-Throughput vLLM Inference Engine
- Deployed self-hosted vLLM inference endpoints utilizing continuous batching and PagedAttention to serve multiple concurrent student sessions.
- Supported 2,000+ concurrent LLM advising interactions with sub-second time-to-first-token (TTFT).
3. Human-in-the-Loop (HITL) Verification
- Incorporated human-in-the-loop validation checkpoints where academic advisors can review, amend, and approve generated degree plans with 1-click workflows.
- Reduced student advising turnaround times by 50% across trial departments.
Publication
- A Recommender System Architecture for University Curriculum Advising
Zerui Ma, Michael Hahsler, Peter Moore.
Proceedings of the AAAI Symposium Series, 5(1), 235–241 (2025).
DOI: 10.1609/aaaiss.v5i1.35593.
