The Operational AI Leap - Zero ETL for Operational AI¶

Overview¶

The Operational AI Leap demonstrates a modern paradigm shift enabling ML engineers to run real-time vector search and generative LLM inference natively inside the database engine. By connecting Google Colab Enterprise directly to live AlloyDB data and joining BigQuery Data Lakes via zero-copy federation, this demo proves how enterprise-grade AI recommendation agents can be deployed in hours with absolutely zero data movement tax.

Zero-ETL Architecture: Eliminating ETL pipelines by connecting ML environments directly to live operational data
In-Database Generative AI: Invoking Gemini LLM endpoints directly inside database SQL via secure IAM integration
Multi-Index Optimization: Fusing Dense Vectors, Sparse Vectors, and Full-Text Search into a single unified plan
Lakehouse Federation: Executing real-time, zero-copy joins between live databases and BigQuery Data Lakes
Compute Isolation: Offloading high-throughput AI workloads onto dynamically scaling Read Pools

The demo proves that Zero-ETL workflows accelerate AI deployment cycles from months to hours while protecting primary application performance.

Demo Architectural Flow Diagram¶

flowchart TD
    subgraph Client ["Client ML Workspace (Colab Enterprise)"]
        ColabNotebook["fa:fa-laptop-code Colab Enterprise Notebook
(init_connection_pool)"]
    end

    subgraph AlloyDB ["AlloyDB for PostgreSQL (Operational Database)"]
        direction TB
        subgraph Extensions ["AI Extension Layer"]
            GoogleML["fa:fa-brain google_ml_integration
(gemini-2.5-flash-lite-global)"]
            BigQueryFDW["fa:fa-network-wired bigquery_fdw
(Server & User Mappings)"]
        end
        subgraph Data ["Operational Schema"]
            LiveProducts["fa:fa-database Products Table
(DML ai.generate)"]
            Indices["fa:fa-bolt Multi-Index Optimizations
(ScaNN, HNSW, and GIN)"]
            LiveProducts -.->|"Index Mappings"| Indices
        end
        subgraph QueryEngine ["Query Planner"]
            RRF["fa:fa-balance-scale SQL Reciprocal Rank Fusion
(Outer Joins & Ranking)"]
            ReadPools["fa:fa-server Scaling Read Pools
(Offloads Analytical Compute)"]
        end

        LiveProducts --> RRF
        BigQueryFDW -->|"Manual Foreign Tables"| RRF
    end

    subgraph DataCloud ["Google Cloud GenAI & Data Lake"]
        VertexAI["fa:fa-cloud Vertex AI API
(Embedding & Generative Completions)"]
        BigQueryLake["fa:fa-database BigQuery Analytics
(Historical order_items Dataset)"]
    end

    ColabNotebook -->|"Direct connection pool (0ms lag)"| LiveProducts
    GoogleML -->|"Natively Call ai.generate()"| VertexAI
    BigQueryFDW -->|"Zero-Copy Federated Join"| BigQueryLake
    RRF -->|"The Magic Moment Context Block"| GoogleML
    VertexAI -->|"Return Personalized Recommendation Text Card"| ColabNotebook
    ReadPools -.->|"Isolates Query Compute"| RRF

    style ColabNotebook fill:#4285F4,stroke:#3c4043,color:#FFFFFF
    style LiveProducts fill:#34A853,stroke:#3c4043,color:#FFFFFF
    style Indices fill:#F8F9FA,stroke:#34A853,color:#202124
    style GoogleML fill:#FBBC04,stroke:#3c4043,color:#202124
    style BigQueryFDW fill:#FBBC04,stroke:#3c4043,color:#202124
    style RRF fill:#4285F4,stroke:#3c4043,color:#FFFFFF
    style ReadPools fill:#F8F9FA,stroke:#4285F4,color:#202124
    style VertexAI fill:#EA4335,stroke:#3c4043,color:#FFFFFF
    style BigQueryLake fill:#34A853,stroke:#3c4043,color:#FFFFFF

Getting Started¶

Note

Are you taking a Qwiklabs lab where the environment is pre-deployed? Please follow QWIKLABS.md instead!

Prerequisites¶

Google Cloud Project with billing enabled.
Google Cloud SDK (gcloud) installed and configured.
Terraform installed.
psql client installed.

Deploy Base Infrastructure via Terraform¶

Authenticate your Google Cloud account:

gcloud auth login
gcloud auth application-default login

Set your active Google Cloud project:

gcloud config set project YOUR_PROJECT_ID

Set optional Terraform environment variables:

Tip

By default, Terraform automatically detects your active GCP project ID and public IP address, and automatically generates a secure 16-character AlloyDB password.

Important

If deploying in an internal Google Argolis environment, set the TF_VAR_argolis flag to true to apply necessary organization policy overrides:

export TF_VAR_argolis="true"

Initialize and apply the Terraform configuration:
```
terraform init
terraform apply
```
When the deployment completes, note the Terraform outputs printed in your terminal:
- demo_app_url: URL of the deployed Cymbal Shops eCommerce application.
- notebook_gcs_uri: GCS path to the companion Colab Enterprise notebook.
- alloydb_password: Generated AlloyDB password.

Explore the Live Storefront Application¶

Open the demo_app_url output URL in your browser to view the live Cymbal Shops eCommerce catalog powered by database-native vector search.

Running the Companion Colab Enterprise Notebook¶

Terraform automatically provisions a private VPC-peered Colab Enterprise Runtime Template named Cymbal Shops Colab Template (stylesearch-colab-template) and uploads the companion notebook to Cloud Storage (notebook_gcs_uri).

To launch the interactive notebook:

In the Google Cloud Console top search bar, search for Colab Enterprise and select Colab Enterprise from the results.
In the left navigation sidebar, click My Notebooks.
Click the Import notebook button at the top of the page.
Under Import source, select Cloud Storage.
In the Cloud Storage file field, click Browse and navigate to your project bucket (or paste your notebook_gcs_uri output, e.g., gs://YOUR_PROJECT_ID/operational-ai-leap.ipynb) and select operational-ai-leap.ipynb.
Click Import at the bottom of the dialog. The notebook will open in your browser.
In the top-right corner of the imported notebook, click the dropdown triangle ▾ next to the connection status indicator (next to the Connect button).
From the dropdown menu, select Change runtime type (or Connect to a runtime).
In the Connect to Agent Platform Runtime panel on the right side:
- Click the Runtimes (or Runtime template) dropdown menu.
- Select Cymbal Shops Colab Template (stylesearch-colab-template).
Verify that the Network and Subnetwork fields show demo-vpc.
Click Connect at the bottom of the panel.
Wait a few seconds for the runtime instance to start and indicate Connected in green.
Once connected, run through the notebook cells sequentially from top to bottom to explore database-native AI and zero-copy Lakehouse federation!

Special thanks¶

I would like to extend special thanks to Paul Ramsey (paulramsey@) for his excellent Cymbal Shops StyleSearch AlloyDB AI Demo which served as the foundation for this demo.

License¶

Please refer to the LICENSE file for details.

Disclaimer¶

This is NOT an officially supported Google product.

This software is provided "as is", without warranty of any kind, expressed or implied, including but not limited to, the warranties of merchantability, fitness for a particular purpose, and/or infringement.

See LICENSE file for additional details.