Understanding Retrieval-Augmented Generation (RAG) on Google Cloud Platform (GCP)

Retrieval-Augmented Generation (RAG) represents a transformative leap in the application of large language models (LLMs). By integrating external knowledge sources, RAG addresses key limitations of LLMs, such as outdated training data and hallucinated responses.

Watch now! > >

In this webinar, Jeroen Overschie, a machine learning engineer at Xebia Data, explains how RAG works, its applications, and its implementation levels on Google Cloud Platform (GCP). 

Jeroen shared, “RAG bridges the gap between static knowledge and real-world, up-to-date information. It’s not just about making AI smarter—it’s about making AI practically useful.” Let’s dive deeper into how RAG can revolutionize data-driven systems.

Speaker 

 

Jeroen Overschie Machine Learning Engineer at Xebia Data

Key Takeaways 

• Dynamic Progression: From basic vector-based searches to advanced multimodal capabilities, RAG evolves with user needs. 
• Seamless Integration: GCP tools offer a cohesive platform to build, manage, and scale RAG systems. 
• Empowered Decision-Making: RAG enhances the practical application of AI by delivering accurate, real-time, and actionable insights. 

Agenda

1. Overview of Retrieval-Augmented Generation (RAG)
2. Why RAG?
3. RAG Levels Overview
4. Level 1: Basic RAG
   (Embeddings, vector search, Cloud Run, Vertex AI)
5. Level 2: Hybrid Search
   (Combining keyword search and vector search, Reciprocal rank fusion)
6. Level 3: Advanced Data Formats
   (Unstructured data, PDF parsing, Document AI)
7. Level 4: Multimodal RAG
   (Multimedia processing, Encoding documents, direct multimodal input)
8. Building RAG Systems with GCP