KEERTHI UDAYAKUMAR
An ambitious Data Scientist at Red Hat with two years of experience specialized in the development and optimization of Large Language Model (LLM) applications. A dedicated open-source contributor focusing on reducing the "token tax" and improving the efficiency of RAG systems within enterprise environments. I was a featured speaker at DevConf.US 2025, where I shared insights on high-volume prompt engineering and structured data optimization.
RED HAT
ASSOCIATE DATA SCIENTIST
Session
We will explore how to transform flat review data into a semantically rich knowledge graph, mapping entities and their multi-hop relationships. By leveraging KGNNs, the system performs sophisticated relational retrieval that standard vector-only RAG misses. We’ll dive into:
Graph Construction: Building a scalable knowledge graph from Yelp's business and review data.
KGNN Integration: How Graph Neural Networks improve the "Retrieval" phase by identifying deep-seated patterns and entity associations.
The RAG Pipeline: Connecting graph-based context to the LLM generation layer for more accurate, fact-grounded recommendations.
Benchmarking: Comparing traditional vector-RAG against KGNN-RAG in terms of reasoning quality and factual consistency.
Attendees will leave with a practical roadmap for implementing Graph-Augmented AI to solve complex, multi-entity retrieval challenges in production.