Production-Ready AI with Google Cloud Learning Path

We're excited to launch the Production-Ready AI with Google Cloud Learning Path, a free series designed to take your AI projects from prototype to production.

This page is the central hub for the curriculum. We'll be updating it weekly with new modules from now through mid-December.

Why We Built This: Bridging the Prototype-to-Production Gap

Generative AI makes it easy to build an impressive prototype. But moving from that proof-of-concept to a secure, scalable, and observable production system is where many projects stall. This is the prototype-to-production gap. It's the challenge of answering hard questions about security, infrastructure, and monitoring for a system that now includes a probabilistic model.

It’s a journey we’ve been on with our own teams at Google Cloud. To solve for this ongoing challenge, we built a comprehensive internal playbook focused on production-grade best practices. After seeing the playbook's success, we knew we had to share it.

This learning path is that playbook, adapted for all developers. The path's curriculum combines the power of Gemini models with production-grade tools like Vertex AI, Google Kubernetes Engine (GKE), and Cloud Run.

We're excited to share this curriculum with the developer community. Share your progress and connect with others on the journey using the hashtag #ProductionReadyAI. Happy learning!

The Curriculum

Developing Apps that use LLMs

Start with the fundamentals of building applications and interacting with models using the Vertex AI SDK.

Summary: Your First AI Application is Easier Than You Think

Go to lab!
Developing LLM Apps with the Vertex AI SDK

Objective: Build a Gemini chatbot with the Vertex AI SDK, integrating real-time data via external tools and refining outputs with prompt engineering.

Deploying Open Models

Learn to serve and scale open source models efficiently by deploying them on production-grade platforms like Google Kubernetes Engine (GKE), Cloud Run, and Vertex AI endpoints.

Summary: Hands-on with Gemma 3 on Google Cloud

Go to labs!

• Serving Gemma 3 with vLLM on Cloud Run
  • Objective: Deploy Gemma 3 to Cloud Run using vLLM, leveraging GPU acceleration to expose an OpenAI-compatible API endpoint.
• Deploying Open Models on GKE
  • Objective: Prototype locally using Ollama, then deploy a scalable inference service to GKE Autopilot using standard Kubernetes manifests.

Developing Agents

Learn to build AI agents that can reason, plan, and use tools to accomplish complex tasks with the Agent Development Kit (ADK).

Summary: Build Your First ADK Agent Workforce

Go to labs!

• Building AI Agents with ADK:The Foundation
  • Objective: Write the essential code to define and run a basic agent using ADK.
• Building AI Agents with ADK:Empowering with Tools
  • Objective: Learn how to use tools to interact with external applications and services.
• Build Multi-Agent Systems with ADK
  • Objective: Orchestrate a complex, automated workflow with a team of specialized agents that work in sequence, in loops, and in parallel.

Securing AI Applications

Master the essential practices for securing your infrastructure, data, and AI-powered endpoints in a production environment.

Summary: Building a Production-Ready AI Security Foundation

Go to labs!

• Securing AI Applications
  • Objective: Learn to use Model Armor to secure Generative AI applications against prompt injection and data leakage.
• Securing Data Used for AI Applications
  • Objective: Build an automated pipeline to inspect, classify, and de-identify PII for use in AI development using Sensitive Data Protection.
• Securing Infrastructure for AI Applications
  • Objective: Secure an AI development environment by implementing network isolation, hardened compute instances, and protected storage.

Deploying Agents

Take your agents to production by deploying them on scalable, managed platforms like Agent Engine, Cloud Run, and Google Kubernetes Engine (GKE).

Summary: From Code to Cloud: Three Labs for Deploying Your AI Agent

Go to labs!

• Deploy ADK Agents to Agent Engine
  • Objective: Deploy a multi-agent system with Agent Engine without provisioning any infrastructure.
• Build and deploy an ADK agent on Cloud Run
  • Objective: Containerize an ADK agent and deploy it to a secure public HTTPS endpoint on Cloud Run to experience the speed of a serverless workflow.
• Deploy ADK agents to Google Kubernetes Engine (GKE)
  • Objective: Deploy an ADK agent to a managed GKE cluster, configuring autoscaling and resource limits using standard manifests.

Evaluation

Discover how to rigorously evaluate the performance of your LLM outputs, agents, and RAG systems to ensure quality and reliability.

Summary: Master Generative AI Evaluation: From Single Prompts to Complex Agents

Go to labs!

• Evaluating Single LLM Outputs With Vertex AI Evaluation
  • Objective: Run your first automated evaluation job to measure the quality of raw LLM responses.
• Evaluate RAG Systems with Vertex AI
  • Objective: Assess the performance of a RAG pipeline by measuring both retrieval quality and generation accuracy.
• Evaluating Agents with ADK
  • Objective: Capture agent execution traces and apply automated evaluation to ensure your agent makes the right decisions.
• Build and Evaluate BigQuery Agents using ADK and GenAI Eval Service
  • Objective: Build a data-driven agent and test its ability to generate accurate SQL and data insights.

Agent Production Patterns

Learn how to enhance your agent's capabilities with agentic RAG, Model Context Protocol (MCP) tools, and Agent to Agent (A2A) protocol.

Summary: Building Connected Agents with MCP and A2A

Go to labs!

• Getting Started with MCP, ADK and A2A
  • Objective: Build a specialized currency ADK agent that leverages MCP and A2A communication.
• MCP Toolbox for Databases: Making BigQuery datasets available to MCP clients
  • Objective: Configure the MCP Toolbox to expose a public BigQuery dataset to an AI agent, enabling it to query and analyze datasets using natural language.
• Build a Travel Agent using MCP Toolbox for Databases and Agent Development Kit (ADK)
  • Objective: Build a full-stack agent that interacts with a Cloud SQL database, demonstrating how to securely expose relational data to an AI agent.

From Data Foundations to Advanced RAG

Learn to build high-performance RAG systems by mastering the full data lifecycle, from generating vector embeddings within your database to implementing advanced retrieval patterns.

The AI Data Layer Foundation

Discover how to transform your operational databases into AI-ready vector stores. Learn to generate embeddings, perform semantic search, and leverage built-in AI functions directly within AlloyDB and Cloud SQL.

Summary: Coming Soon!

Go to labs!

• Getting started with Vector Embeddings with AlloyDB AI
  • Objective: Generate embeddings and perform semantic search within AlloyDB to ground Gen AI responses.
• Getting started with Vector Embeddings in Cloud SQL for PostgreSQL
  • Objective: Generate embeddings and perform semantic search within Cloud SQL for PostgreSQL to ground Gen AI responses.
• Getting started with Vector Embeddings in Cloud SQL for MySQL
  • Objective: Generate embeddings and perform semantic search within Cloud SQL for MySQL to ground Gen AI responses.
• Multimodal Embeddings in AlloyDB
  • Objective: Perform semantic search on both text and images using multimodal embeddings in AlloyDB.
• AlloyDB AI Operators and Reranking
  • Objective: Filter data with natural language and apply reranking to improve search precision using AlloyDB AI functions.
• Generate SQL using AlloyDB AI natural language
  • Objective: Generate reliable SQL queries from natural language by configuring custom schema context in AlloyDB.

Building the RAG application

Move beyond basic vector search. Learn to architect robust RAG applications by using advanced retrieval strategies and leveraging tools.

Summary: Coming soon!

Go to labs!

• Intro to Agentic RAG
  • Objective: Build a multi-tool agent that combines retrieval from unstructured documents and structured data to answer reasoning-heavy questions.
• AlloyDB Agentic RAG Application with MCP Toolbox
  • Objective: Deploy the MCP Toolbox to connect an interactive AI application to an AlloyDB database for grounded responses.
• Advanced RAG Techniques
  • Objective: Implement and evaluate advanced strategies (Chunking, Reranking, Query Transformation) to improve the precision and recall of your RAG pipeline.

Fine-Tuning

Go beyond prompting and learn how to fine-tune both open and proprietary models to improve performance on specific tasks.

Summary: Coming soon!

Go to labs!

• Fine-tune Gemini on Vertex AI
  • Objective: Perform supervised fine-tuning on Gemini 2.5 Flash to adapt it for specific tasks like summarization.
• Fine-tune Open Source LLMs on Google Cloud
  • Objective: Build a production-grade fine-tuning pipeline for Llama 2 on GKE using LoRA and PyTorch.

We're committed to making this a living, evolving resource and will be adding to it over time.

Do you feel something is missing? Tell us here!