Skip to content

Zero ETL for Operational AI: The Operational AI Leap

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 zero data movement overhead.

  • Zero-ETL Architecture : Eliminates ETL pipelines by connecting ML environments directly to live operational data.
  • In-Database Generative AI : Invokes Gemini LLM endpoints directly inside database SQL via secure IAM integration.
  • Multi-Index Optimization : Fuses Dense Vectors, Sparse Vectors, and Full-Text Search into a single unified plan.
  • Lakehouse Federation : Executes real-time, zero-copy joins between live databases and BigQuery Data Lakes.
  • Compute Isolation : Offloads 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.

What You'll Need

To complete this lab, you will need:

  • Access to a standard internet browser ( Chrome browser recommended ).
  • Time : Note the lab's Completion time in Qwiklabs, which is an estimate of the time it should take to complete all steps. Plan your schedule so you have time to complete the lab. Once you start the lab, you will not be able to pause and return later because you begin at step 1 every time you start a lab.
  • You do NOT need a Google Cloud Platform account or project. An account, project, and associated resources are provided to you as part of this lab.
  • If you already have your own GCP account, make sure you do not use it for this lab.
  • If your lab prompts you to log into the console, use only the student account provided to you by the lab. This prevents you from incurring charges for lab activities in your personal GCP account.
  • Use a new Incognito window in Chrome or another browser for the Qwiklabs session. Alternatively, you can log out of all other Google and Gmail accounts before beginning the lab.

Start Your Lab

When you are ready, click Start Lab . You can track your lab's progress with the status bar at the top of your screen.

Important: What is happening during this time? Your lab is spinning up GCP resources for you behind the scenes, including an account, a project, resources within the project, and permissions for you to control the resources you will need to run the lab. This means that instead of spending time manually setting up a project and building resources from scratch as part of your lab, you can begin learning more quickly once provisioning completes.

Sign In to Google Cloud Console

Locate Your Lab's GCP Username and Password

To access the resources and console for this lab, locate the Connection Details panel in Qwiklabs. Here you will find the account ID and password for the account you will use to log in to the Google Cloud Platform . If your lab provides other resource identifiers or connection-related information, it will appear on this panel as well.

Access the Console in Incognito Mode

  1. Using the Qwiklabs browser tab/window, preferably in Incognito mode , or the separate browser you are using for the Qwiklabs session, copy the Username from the Connection Details panel and click Open Google Console .
  2. Paste in the Username and then the Password as prompted.
  3. Accept the terms and conditions.
  4. Because this is a temporary account, which you will only have access to for this one lab, do not add recovery options and do not sign up for free trials.

Note: You can view the menu with a list of GCP Products and Services by clicking the navigation menu button at the top-left next to Google Cloud Platform.

Lab Environment Readiness

Welcome to The Operational AI Leap Qwiklabs environment!

In this lab, all foundational infrastructure and data readiness tasks have been automatically provisioned and verified upon starting the lab session. You do not need to run any Terraform commands, data pipelines, or database population scripts manually.

  • AlloyDB for PostgreSQL : Pre-loaded with a live eCommerce product catalog containing over 1,000 verified records and fully optimized vector search indexes including ScaNN , HNSW , and GIN built in parallel—ready for instant Zero-ETL querying.
  • Cymbal Shops Demo Application : A live web storefront application hosted on Cloud Run.
  • BigQuery Data Lake : Cross-region federated copy of historical order and transaction data under dataset thelook_ecommerce .
  • Colab Enterprise Runtime Template : Pre-created and peered securely to the private database VPC named demo-vpc .

Step 1: Locate Your Lab Credentials & Outputs

Step 1.1: Locate Student Visible Outputs

Once you have signed into the Google Cloud Console using your temporary credentials, return briefly to your Qwiklabs instruction panel on the left side of your screen and locate the section titled Student Visible Outputs ( or lab details ).

Step 1.2: Required Lab Outputs Checklist

Make a note of the following exact outputs from the Student Visible Outputs panel:

  • Demo Application URL ( demo_app_url ) : The web URL to view the live Cymbal Shops demo storefront.
  • Colab Notebook GCS URI ( notebook*gcs_uri ) : The Cloud Storage path where your interactive English lab notebook is stored ( _e.g., gs://YOUR_PROJECT_ID/operational-ai-leap.ipynb* ).
  • Colab Notebook GCS URI (KR) ( notebook*gcs_uri_ko ) : The Cloud Storage path for the Korean translated lab notebook ( _e.g., gs://YOUR_PROJECT_ID/operational-ai-leap-ko.ipynb* ).
  • AlloyDB Password ( alloydb_password ) : The generated secure password for the AlloyDB database connection pool.
  • GCP Project ID ( gcp_project_id ) : Your active Google Cloud project ID.
  • GCP Region ( gcp*region ) : The exact Google Cloud region where your lab resources and runtime template were deployed ( _e.g., us-central1 or us-west1* ).
  • AlloyDB Cluster Name ( alloydb_cluster_name ) : Pre-configured cluster name which is stylesearch-cluster .
  • VPC Network Name ( vpc_name ) : Pre-configured private network which is demo-vpc .
  • Colab Runtime Template ID ( colab_runtime_template_id ) : Pre-configured Agent Platform runtime template which is stylesearch-colab-template .

Step 2: Explore the Live Storefront Application

  1. Click or copy the Demo Application URL from your Student Visible Outputs into your browser.
  2. Explore the Cymbal Shops eCommerce catalog. Notice how semantic vector search and generative AI recommendations power product discovery right out of the box without any external ETL replication delays.

Step 3: Run the Companion Interactive Notebook

To dive deep into the database-native AI architecture, you will run an interactive Jupyter notebook inside Google Colab Enterprise . You may choose either the English version ( operational-ai-leap.ipynb ) or the Korean version ( operational-ai-leap-ko.ipynb ).

Step 3A: Import the Notebook into Colab Enterprise

  1. In the Google Cloud Console top search bar, search for Colab Enterprise and select Colab Enterprise from the results.
  2. In the top-right corner of the Colab Enterprise page, check the Region dropdown selector. Ensure it is set to your exact GCP Region ( gcp*region ) from your Student Visible Outputs ( _e.g., us-central1 or us-west1* ). If it shows a different region, click the dropdown and select your matching GCP Region so your runtime and private network align.
  3. In the left navigation sidebar, click My Notebooks .
  4. Click the Import notebook button at the top of the page.
  5. Under Import source , select Cloud Storage .
  6. In the Cloud Storage file field, click Browse and navigate into your project bucket using your notebook_gcs_uri or notebook_gcs_uri_ko output path, then select your preferred notebook file: operational-ai-leap.ipynb for English or operational-ai-leap-ko.ipynb for Korean.
  7. Click Import at the bottom of the dialog. The notebook will open directly in your browser.

Step 3B: Connect to the Pre-Provisioned Runtime Template

To securely connect your notebook to the private AlloyDB database instance, you must connect to the pre-created VPC-peered runtime template in your matching region:

  1. In the top-right corner of your imported notebook, click the dropdown arrow ▾ next to the Connect button.
  2. From the dropdown menu, select Change runtime type or Connect to a runtime .
  3. In the Connect to Agent Platform Runtime panel on the right side, verify or select your matching GCP Region ( gcp_region ) if prompted.
  4. Click the Runtimes or Runtime template dropdown menu.
  5. Select Cymbal Shops Colab Template ( stylesearch-colab-template ).
  6. Verify that the Network field shows demo-vpc and Subnetwork shows demo-vpc-subnet .
  7. Click Connect at the bottom of the panel.
  8. Wait a few seconds for the runtime instance to start and indicate Connected in green.

Step 3C: Execute the Lab Workflows

Once connected, and before clicking run on all cells , follow these steps to configure and execute your pipeline:

Step 3C.1: Configure Connection Variables ( Required First Step )

  1. Locate the code cell directly under the heading Connection Setup & Environment Configuration ( around Cell 5 ).
  2. Update project_id = "your-project-id" with your exact GCP Project ID ( gcp_project_id ) from your Student Visible Outputs.
  3. Update alloydb_password = "your-db-password" with your exact AlloyDB Password ( alloydb_password ) from your Student Visible Outputs.
  4. Run this cell by clicking the play button ▶ or pressing Shift + Enter .

Step 3C.2: Execute Remaining Pipeline Stages

  1. Database Connection Pool : Run the next cell to initialize low-latency asyncpg connections to AlloyDB with 0ms replication lag.
  2. In-Database Generative AI : Execute SQL queries invoking ai.generate() and embedding() natively inside PostgreSQL using extension google_ml_integration .
  3. Reciprocal Rank Fusion (RRF) : Combine vector similarity, full-text keyword search via index GIN , and sparse BM25 ranking into unified recommendation cards.
  4. Zero-Copy Lakehouse Federation : Run live federated joins using bigquery*fdw across BigQuery historical order datasets _without ETL pipelines*.

End Your Lab

When you have completed your lab, click End . Qwiklabs removes the resources you have used and cleans the account for you.

You will be given an opportunity to rate the lab experience. Select the applicable number of stars, type a comment, and then click Submit .

Note: The number of stars indicates the following:

  • 1 star = Very dissatisfied
  • 2 stars = Dissatisfied
  • 3 stars = Neutral
  • 4 stars = Satisfied
  • 5 stars = Very satisfied

You may close the dialog if you do not want to provide feedback.

Additional Resources

  • For more information about Google Cloud Training and Certification, see https://cloud.google.com/training/
  • For more Google Cloud Platform Self-Paced Labs, see http://run.qwiklabs.com
  • For feedback, suggestions, or corrections, please use the Support tab.