Vietnamese Labor Law AI Assistant

This project explores how AI can help users search and understand Vietnamese labor-law documents in a controlled, citation-aware way. The assistant retrieves legal context, expands it through a legal knowledge graph, and validates citations before producing a grounded answer.

View on GitHub

Retrieved context

Graph relation

Citation check

Problem

Vietnamese labor-law information is fragmented across official documents and related legal structures. A natural-language question needs an answer grounded in retrieved legal context, not unsupported model generation.

Solution

The project uses a scoped legal corpus, hybrid retrieval, legal graph expansion, grounded answer generation, and deterministic citation validation to produce legal-information responses.

Outcome / What this demonstrates

The project demonstrates how retrieval, graph context, and citation checks can make an AI assistant more grounded, transparent, and suitable for legal-information research.

Highlights

Scoped Vietnamese labor-law corpus

Hybrid dense/sparse retrieval with Qdrant

Legal graph expansion with Neo4j

Grounded answer generation

Deterministic citation validation

Evaluation and reproducibility pipeline

Tech stack

PythonFastAPINext.jsTypeScriptQdrantNeo4jRAGGraphRAGLLM

This project is a legal-information assistant and research project. It is not a replacement for professional legal advice.