Social Graph of Prominent Tech figures using GraphRAG on Relational AI

overview

In this example architecture, you will build a social graph of relationships between prominent tech figures from a text corpus consisting of their CVs. You will also learn how to extract named entities and relationships from the corpus using Snowflake Cortex and Llama 3.0, and to build a GraphRAG using the RelationalAI model. 

  • Install the RelationalAI Native App from Snowflake Marketplace
  • Load the corpus of CVs from Azure Blob Store into a Snowflake Table
  • Extract named entities and relationships from the corpus using Snowflake Cortex and Llama 3.0
  • Build a RelationalAI model and GraphRAG to answer questions about the relationships from the corpus

Snowflake Features Used
Quickstart
Build Now Getting started with GraphRAG and RelationalAI

Solution Architecture: Data Engineering Pipelines with Snowflake Notebooks

  • Create the compute pool to host the RAI engine
  • Load the corpus from Azure blob into Snowflake table
  • Create UDFs for entity and relationship extraction using Llama 3.0 on Cortex AI
  • Synchronize the source node and edges tables from Snowflake to RAI
  • Build a graph model for community detection and summarization

HANDS-ON TECHNICAL RESOURCES