Recap and next steps

In this chapter, we combined the majority of the important concepts that we’ve covered throughout this book and explained how they all tie together. We did this by building solutions that incorporated many of the topics and Google Cloud products we’ve discussed.

First, you built a RAG implementation, piece by piece, focusing on the combination of various generative AI concepts in that process, such as using Google Cloud Document AI to break a document into chunks in a manner that preserves the hierarchical structure of the original document. This solution also included using the Google Cloud text embedding LLM to create embeddings for the document chunks and using Vertex AI Vector Search to store and index those embeddings. You used the resulting solution to implement a question-answering use case with Gemini, grounding the answers in the contents of the document. To do this, you used the Google Cloud text embedding LLM to create embeddings of...