This repository provides code samples, blueprints, and guides for leveraging Generative AI on Google Cloud's Vertex AI platform. It targets engineers and researchers looking to implement advanced AI solutions, offering practical examples for foundation models, RAG, agents, and operationalizing research models.

How It Works

The project showcases practical applications of Google Cloud's AI/ML infrastructure and Vertex AI services. It features code samples and hands-on labs demonstrating the use of foundation models, Retrieval Augmented Generation (RAG), agents, and the operationalization of research models. The approach emphasizes best practices and prescriptive guidance for building and deploying AI workloads.

Quick Start & Requirements

• Install: Primarily Python-based, requiring installation of relevant Google Cloud client libraries and Vertex AI SDK. Specific setup commands depend on the chosen sample.
• Prerequisites: Google Cloud account, Vertex AI API enabled, potentially specific Python versions (e.g., 3.9+), and authentication via gcloud CLI or service accounts. Some samples may require large datasets or specific model access.
• Resources: Setup complexity varies by sample; some may require significant compute resources for training or inference.
• Documentation: https://googlecloudplatform.github.io/applied-ai-engineering-samples/

Highlighted Details

• Comprehensive coverage of Generative AI on Vertex AI, including Gemini, RAG, and Agents.
• Solutions Catalog features practical applications like Open Data Q&A, GenAI for Marketing, and Customer Experience.
• Includes "RAG Playground" for experimenting with Retrieval Augmented Generation techniques using LangChain and Vertex AI.
• Demonstrates operationalization of models from Google DeepMind and Research teams.

Maintenance & Community

Developed by the Google Cloud Applied AI Engineering team. Issues and problems can be reported via GitHub issues.

Licensing & Compatibility

The code is provided for demonstrative purposes and is not an officially supported Google product. The specific license is not explicitly stated in the README, but as a Google Cloud project, it's likely to be permissive for use with Google Cloud services.

Limitations & Caveats

This repository contains code for demonstrative purposes only and is not an officially supported Google product. Users should be aware of potential changes and the non-production nature of the samples.