LLMs, Vectors, Graph Databases and RAG in the Cloud

​We are super excited to organize an evening event on LLMs, Graph Databases, and RAG on Google Cloud Platform. Come and join us for this in-person meetup in Reston, VA. Pizza will be provided.

​You will learn about RAG from two veteran speakers who have been presenting about it for the past year: Brian Snyder, Google and Sydney Beckett, Neo4j.

​Agenda
4:00 pm: Open doors & mingling
4:45 pm: Intro and Talks
6:30 pm: Closing

​Talk #1: Augmenting and Grounding LLMs with Information Retrieval - Brian Snyder, Google

​Grounding helps LLMs access additional information beyond their training data to improve responses. Explore Retrieval Augmented Generation (RAG), a method of enriching prompts with supporting data stored in vector databases. Discover how this approach improves response quality and context, reducing the necessity for model retraining.

​Brian has been a partner engineer at Google for 5 years.  In that time, he's worked with many different partners across a few different specialties.  In 2024, when the GenAI COE team was formed, Brian joined with the mission of helping to build deep GenAI expertise across the partner ecosystem and to accelerate GenAI adoption

​Talk #2: Going Beyond Vectors - Sydney Beckett

​Retrieval Augmented Generation (RAG) with vector embeddings are enormously powerful. But there are also limitations to vectors and general challenges to implementing RAG systems. In this talk, we'll discuss some of those challenges and introduce using stored knowledge graphs to address some of them. We showcase how to integrate an LLM to query the newly-created knowledge graph with plain English questions. This will leverage Neo4j’s graph database along with GCP’s VertexAI.

​Sydney “Syd” became a graph enthusiast through her work with clients to build graph-based solutions and support data science teams when she was a consultant. Now she uses her graph expertise to help users realize the value of graph technology for their organization.  She also contributes by teaching Neo4j graph database and data science training. Syd’s hobbies include interior design and defeating the car navigation system’s estimated drive time.