← Back Home
CuRAG — Knowledge Graph Curation System
CuRAG, a project under supervision of Prof. Dr. Ralph Ewerth,
Multimodal Modelling & Machine Learning (M3L) Group ,
is a research driven knowledge graph curation platform designed
to support domain experts in validating, managing, and exploring
scientific knowledge extracted from research documents.
The system integrates Neo4j-based graph structures with
Retrieval Augmented Generation (RAG) workflows for faster query answering and local LLMs for intelligent
suggestions.
Responsibilities
Designed and developed interactive interfaces enabling domain experts
to curate and validate Neo4j based knowledge graphs.
Integrated question answering workflows grounded in attached PDF documents
using local Large Language Models (LLMs).
Implemented workflows for creating, modifying,
and managing nodes, relationships, and metadata.
Researched and worked with information extraction pipeline to extract the entities from the research papers.
Technologies & Domains
React
Python
Neo4j
Knowledge Graphs
RAG Systems
Local LLMs
Full Stack Development
Graph Databases
Information Extraction
Scientific Document Processing
Detailed documentation and source code are currently private
due to ongoing research and development work.