The Architecture Center provides content resources across a wide variety of AI and machine learning subjects. This page provides information to help you get started with generative AI, traditional AI, and machine learning. It also provides a list of all the AI and machine learning (ML) content in the Architecture Center.
Get started
The documents listed on this page can help you get started with designing, building, and deploying AI and ML solutions on Google Cloud.
Explore generative AI
Start by learning about the fundamentals of generative AI on Google Cloud, on the Cloud documentation site:
- To learn the stages of developing a generative AI application and explore the products and tools for your use case, see Build a generative AI application on Google Cloud.
- To identify when generative AI, traditional AI (which includes prediction and classification), or a combination of both might suit your business use case, see When to use generative AI or traditional AI.
- To define an AI business use case with a business value-driven decision approach, see Evaluate and define your generative AI business use case.
- To address the challenges of model selection, evaluation, tuning, and development, see Develop a generative AI application.
To explore a generative AI and machine learning blueprint that deploys a pipeline for creating AI models, see Build and deploy generative AI and machine learning models in an enterprise. The guide explains the entire AI development lifecycle, from preliminary data exploration and experimentation through model training, deployment, and monitoring.
Browse the following example architectures that use generative AI:
- Generative AI document summarization
- Generative AI knowledge base
- Generative AI RAG with Cloud SQL
- Infrastructure for a RAG-capable generative AI application using GKE
- Infrastructure for a RAG-capable generative AI application using Vertex AI
- Model development and data labeling with Google Cloud and Labelbox
For information about Google Cloud generative AI offerings, see Vertex AI, Gemini API, and running your foundation model on GKE.
Design and build
To select the best combination of storage options for your AI workload, see Design storage for AI and ML workloads in Google Cloud.
Google Cloud provides a suite of AI and machine learning services to help you summarize documents with generative AI, build image processing pipelines, and innovate with generative AI solutions.
Keep exploring
The documents that are listed later on this page and in the left navigation can help you build an AI or ML solution. The documents are organized in the following categories:
- Generative AI: Follow these architectures to design and build generative AI solutions.
- Model training: Implement machine learning, federated learning, and personalized intelligent experiences.
- MLOps: Implement and automate continuous integration, continuous delivery, and continuous training for machine learning systems.
- AI and ML applications: Build applications on Google Cloud that are customized for your AI and ML workloads.
AI and machine learning resources in the Architecture Center
You can filter the following list of AI and machine learning resources by typing a product name or a phrase that's in the resource title or description.
|
Architecture for MLOps using TensorFlow Extended, Vertex AI Pipelines, and Cloud Build This document describes the overall architecture of a machine learning (ML) system using TensorFlow Extended (TFX) libraries. It also discusses how to set up a continuous integration (CI), continuous delivery (CD), and continuous training (CT) for... Products used: Cloud Build |
|
Best practices for implementing machine learning on Google Cloud Introduces best practices for implementing machine learning (ML) on Google Cloud, with a focus on custom-trained models based on your data and code. Products used: Vertex AI, Vertex AI, Vertex AI, Vertex Explainable AI, Vertex Feature Store, Vertex Pipelines, Vertex Tensorboard |
|
Build and deploy generative AI and machine learning models in an enterprise Describes the generative AI and machine learning (ML) blueprint, which deploys a pipeline for creating AI models. |
|
Building an ML vision analytics solution with Dataflow and Cloud Vision API How to deploy a Dataflow pipeline to process large-scale image files with Cloud Vision. Dataflow stores the results in BigQuery so that you can use them to train BigQuery ML pre-built models. Products used: BigQuery, Cloud Build, Cloud Pub/Sub, Cloud Storage, Cloud Vision, Dataflow |
|
Cross-silo and cross-device federated learning on Google Cloud Provides guidance to help you create a federated learning platform that supports either a cross-silo or cross-device architecture. |
|
Data science with R on Google Cloud: Exploratory data analysis Shows you how to get started with data science at scale with R on Google Cloud. This document is intended for those who have some experience with R and with Jupyter notebooks, and who are comfortable with SQL. Products used: BigQuery, Cloud Storage, Notebooks |
|
Design storage for AI and ML workloads in Google Cloud Map the AI and ML workload stages to Google Cloud storage options, and select the recommended storage options for your AI and ML workloads. Products used: Cloud Storage, Filestore, Persistent Disk |
|
Geospatial analytics architecture Learn about Google Cloud geospatial capabilities and how you can use these capabilities in your geospatial analytics applications. Products used: BigQuery, Dataflow |
|
Google Workspace Backup with Afi.ai Describes how to set up an automated Google Workspace backup using Afi.ai. Products used: Cloud Storage |
|
Guidelines for developing high-quality ML solutions Collates some guidelines to help you assess, ensure, and control quality in machine learning (ML) solutions. |
|
Image processing using microservices and asynchronous messaging Shows how to implement microservices using Google Kubernetes Engine (GKE) and Pub/Sub to invoke long-running processes asynchronously. Products used: Artifact Registry, Cloud Build, Cloud Pub/Sub, Cloud SQL, Cloud Storage, Cloud Vision, Compute Engine, Container Registry, Google Kubernetes Engine (GKE) |
|
Infrastructure for a RAG-capable generative AI application using GKE Shows you how to design the infrastructure for a generative AI application with RAG using GKE. Products used: Cloud SQL, Cloud Storage, Google Kubernetes Engine (GKE) |
|
Infrastructure for a RAG-capable generative AI application using Vertex AI Design infrastructure to run a generative AI application with retrieval-augmented generation. Products used: BigQuery, Cloud Logging, Cloud Monitoring, Cloud Pub/Sub, Cloud Run, Cloud Storage, Document AI, Vertex AI |
|
Intelligent Products Essentials reference architecture Describes a reference architecture for technical decision makers who want to connect devices and build Cloud IoT apps on Google Cloud using Intelligent Products Essentials. Products used: BigQuery, Cloud Pub/Sub, Firebase, Intelligent Products Essentials, Looker, Run, Scheduler |
|
Jump Start Solution: AI/ML image processing on Cloud Functions Analyze images by using pretrained machine learning models and an image-processing app deployed on Cloud Functions. |
|
Jump Start Solution: Analytics lakehouse Unify data lakes and data warehouses by creating an analytics lakehouse using BigQuery to store, process, analyze, and activate data. |
|
Jump Start Solution: Data warehouse with BigQuery Build a data warehouse with a dashboard and visualization tool using BigQuery. |
|
Jump Start Solution: Generative AI document summarization Process and summarize documents on demand by using Vertex AI Generative AI and large language models (LLMs). |
|
Jump Start Solution: Generative AI Knowledge Base Extract question and answer pairs from documents on demand by using Vertex AI Generative AI and large language models (LLMs)... |
|
Jump Start Solution: Generative AI RAG with Cloud SQL Deploy a retrieval augmented generation (RAG) application with vector embeddings and Cloud SQL. |
|
Minimizing real-time prediction serving latency in machine learning Explores common architectures on Google Cloud for providing predictions from machine learning models, as well as techniques for minimizing the prediction serving latency of ML systems. Products used: AI Platform, Bigtable, Cloud Memorystore, Cloud Pub/Sub, Dataflow, Datastore, Google Kubernetes Engine (GKE) |
|
MLOps with Intelligent Products Essentials Describes a reference architecture for implementing MLOps using Intelligent Products Essentials and Vertex AI. Products used: BigQuery, Cloud Scheduler, Dataflow, Dataproc, Intelligent Products Essentials, Vertex AI |
|
MLOps: Continuous delivery and automation pipelines in machine learning Discusses techniques for implementing and automating continuous integration (CI), continuous delivery (CD), and continuous training (CT) for machine learning (ML) systems. |
|
Model development and data labeling with Google Cloud and Labelbox Provides guidance for building a standardized pipeline to help accelerate the development of ML models. |
|
Monitoring time-series data with OpenTSDB on Bigtable and GKE Time-series data is a highly valuable asset that you can use for various applications, including trending, monitoring, and machine learning. You can generate time-series data from server infrastructure, application code, and other sources. OpenTSDB... Products used: Bigtable, Google Kubernetes Engine (GKE) |
|
Protecting confidential data in Vertex AI Workbench user-managed notebooks Describes controls and layers of security that you can use to help protect confidential data in Vertex AI Workbench user-managed notebooks. Products used: BigQuery, Notebooks |
|
Reduce your Google Cloud carbon footprint Explains the Google Cloud approach to environmental sustainability. It includes information and other resources that you can use to understand your carbon footprint on Google Cloud. Products used: BigQuery, Compute Engine, Google Kubernetes Engine (GKE) |
|
Scalable TensorFlow inference system Describes how to design and deploy a high performance online inference system for deep learning models by using an NVIDIA® T4 GPU and Triton Inference Server. Products used: Cloud Load Balancing, Google Kubernetes Engine (GKE) |
|
Serve Spark ML models using Vertex AI Shows how to serve (run) online predictions from machine learning (ML) models that are built by using Spark MLlib and managed by using Vertex AI. Products used: Vertex AI |
|
Use Kubeflow Pipelines for propensity modeling on Google Cloud Describes an example of an automated pipeline in Google Cloud that performs propensity modeling. Products used: BigQuery, Cloud Functions, Vertex AI |