TURNING UNSTRUCTURED DOCUMENTS INTO A QUERYABLE ENTITY NETWORK.
Notes, papers, and code docs converted into a persistent graph with community detection and path search.
ABOUT.
RoleFULL-STACK DEVELOPER
Duration6 MONTHS
StatusIN DEVELOPMENT
StackNODE.JS / TYPESCRIPT / NEO4J / REACT
PROBLEM.
Personal research scattered across documents, notes, and browser tabs cannot be queried. Connections between sources exist but stay invisible without manual cross-referencing.
SOLUTION.
An ingestion pipeline converts arbitrary inputs into typed entities and relations stored in a graph database. Community detection clusters the graph, and a path-search API answers 'how are these two concepts related' queries.
RESULTS.
The graph persisted through multiple pipeline rewrites without data loss. Retrieval by entity relation replaced keyword search as the primary way to explore the knowledge base.
