Course Overview: AI Data Agent Development

Orchestrate Workflows with the Data Agent Kit is an advanced-level, instructor-led course designed for practitioners who want to bring natural-language and AI-driven workflows into modern data engineering environments. The course introduces the foundations of agentic data engineering and the Data Agent Kit architecture. Participants learn how ADK-based data agents can interact with Google Cloud data services and how the Model Context Protocol (MCP) provides connectivity between agents and data platforms. Learners build and configure a data agent environment, connect to BigQuery, Spanner, and Cloud SQL, and use the agent to discover and understand enterprise data. The course then explores natural-language-driven data engineering tasks including catalog browsing, universal search, schema and lineage inspection, data quality analysis, SQL transformation generation, BigQuery DataFrames, notebooks, and medallion architecture. Participants also learn how to use the data agent to orchestrate pipelines, build DAGs using natural language, schedule deployments, monitor pipelines, and diagnose failures. The course extends into machine learning workflows using BigQuery ML, including model training, inference, and drift monitoring. Finally, learners explore enterprise security practices including least privilege, principal access boundaries, IAM deny policies, prompt injection protection, document guardrails, and Model Armor.

After completing the AI Data Agent course, participants will be able to:

  • Explain the foundations of agentic data engineering.
  • Describe the architecture of the Data Agent Kit.
  • Explain how the Data Agent Kit integrates with the Agent Development Kit.
  • Build a basic ADK-based data agent.
  • Configure the Data Agent Kit development environment.
  • Configure authentication and MCP connectivity.
  • Apply service account impersonation for secure access.
  • Connect data agents to BigQuery, Spanner, and Cloud SQL.
  • Discover and explore enterprise data using natural-language interactions.
  • Browse data catalogs using the Data Agent Kit.
  • Perform universal search across available data resources
  • Inspect schemas, lineage, and data quality using natural-language interactions.
  • Generate data transformations for business data pipelines.
  • Organize data using the medallion architecture.
  • Generate SQL transformations using natural-language prompts.
  • Work with BigQuery DataFrames and notebooks.
  • Orchestrate data pipelines using the data agent.
  • Build DAGs using natural language.
  • Schedule and deploy pipelines using GitHub Actions.
  • Monitor pipeline execution through the agent.
  • Resolve data pipeline incidents.
  • Trace data lineage to identify reconciliation gaps.
  • Diagnose pipeline failures using agent-driven workflows.
  • Build and train models using BigQuery ML.
  • Perform inference and monitor model drift.
  • Secure agent environments using enterprise security controls.
  • Apply least-privilege access using the MCP Tool User role.
  • Configure principal access boundaries.
  • Apply IAM deny policies for MCP.
  • Protect agents against prompt injection.
  • Apply document guardrails.
  • Configure Model Armor to support secure agent workflows.

Upcoming Batches

Loading Dates...

Key Features: AI Data Agent Learning

  • Foundations of Agentic Data Engineering

  • Data Agent Kit Architecture 

  • ADK-Based Data Agent Development 

  • MCP Connectivity and Authentication 

  • BigQuery, Spanner, and Cloud SQL Integration

  • Natural-Language Data Discovery 

  • Data Catalog, Schema, Lineage, and Quality Analysis 

  • AI-Driven Data Transformation 

  • Natural-Language Pipeline Orchestration

  • Pipeline Monitoring and Troubleshooting 

  • BigQuery ML Model Development 

  • Enterprise Agent Security and Governance 

Who should attend this AI Data Agent Certification Training

  • Data Engineers
  • Data Scientists
  • Software Developers and Engineers
  • AI/ML Engineers
  • Cloud Engineers
  • Platform Engineers
  • DevOps Engineers
  • MLOps Engineers
  • Technical Architects
  • Solutions Architects
  • Data Architects
  • Professionals building AI-driven data engineering workflows

Prerequisites of AI Data Agent Course

• Familiarity with Google Cloud data services is recommended • Familiarity with data engineering concepts is beneficial. • Familiarity with SQL and data transformation concepts is recommended. • Basic understanding of the Agent Development Kit (ADK) is beneficial. • Familiarity with MCP concepts is helpful. • Experience with IDE and CLI-based development workflows is recommended. • Basic understanding of BigQuery is beneficial. • Familiarity with cloud security and IAM concepts is recommended.

Why choose CloudThat as your training partner?

  • Specialized Google Cloud and Data Engineering Expertise 

    CloudThat specializes in Google Cloud, data engineering, Generative AI, and agentic application technologies, delivering focused training programs aligned with enterprise data and AI use cases.

  • Industry-Recognized Trainers 

    Our trainers are certified Google Cloud professionals with expertise in BigQuery, data engineering, ADK, AI agents, MCP, cloud security, machine learning, and enterprise architecture.

  • Hands-On Learning Approach 

    CloudThat emphasizes practical learning through guided data agent development, MCP connectivity, data discovery, transformation generation, pipeline orchestration, troubleshooting, BigQuery ML, and security exercises.

  • Customized Learning Paths 

    Training programs are designed for data engineers, cloud professionals, AI/ML engineers, developers, platform engineers, and technical architects working with modern data and AI platforms. 

  • Interactive Learning Experience 

    Sessions include technical demonstrations, hands-on agent development, natural-language data engineering scenarios, architecture walkthroughs, troubleshooting activities, security scenarios, and knowledge checks. 

  • Continuous Learning and Updates 

    Course content is continuously updated to align with advancements in Google Cloud data services, ADK, Data Agent Kit, MCP, BigQuery ML, agentic data engineering, and enterprise AI security. 

Learning Objective of AI Data Agent Development Certifications

  • This course enables learners to use the Data Agent Kit to bring natural-language, AI-driven workflows into modern data engineering environments. Learners will develop the ability to build ADK-based data agents, configure MCP connectivity, connect to Google Cloud data services, discover and analyze enterprise data, generate transformations, and work with BigQuery DataFrames and notebooks.  Learners will also develop the ability to orchestrate and monitor pipelines, build DAGs using natural language, troubleshoot pipeline incidents, perform data lineage analysis, build machine learning models using BigQuery ML, and monitor model drift.  The course further enables learners to apply enterprise security controls including least privilege, principal access boundaries, IAM deny policies, prompt injection protection, document guardrails, and Model Armor.

Course Outline: AI Data Agent Development with Data Agent Kit Download Course Outline

  • The role of agentic AI in modern data engineering
  • Scenario setup with Cymbal Superstore and the Cymbal discount initiative
  • Data engineering agent interfaces

Demo:

  • Data engineering agent interfaces and ADK patterns

Scenario walkthrough:

  • Cymbal Superstore, with architecture walkthrough and knowledge check.

  • Building and installing the development environment
  • Building a basic ADK data agent
  • Installing the Data Agent Kit starter pack
  • Configuring the Data Agent Kit VS Code extension

Hands-on Lab:

  • Build and install the environment and create a basic ADK data agent.

Demo:

  • Data Agent Kit starter pack and VS Code extension configuration.

  • Authentication for data agent workflows and establishing MCP connectivity
  • Service account impersonation
  • Local and remote MCP servers
  • How MCP enables agents to interact with data services

Hands-on Lab:

  • Service account impersonation.

Guided activity:

  • Connect data agents to BigQuery, Spanner, and Cloud SQL

  • Discovering Cymbal Superstore data resources
  • Browsing the data catalog with the data agent
  • Universal search across available data resources
  • Chat-driven schema, lineage, and data quality inspection

Hands-on Lab:

  • Schema and lineage inspection.

Demo:

  • Universal search, catalog browsing, and data quality analysis

  • Generating transformations for the discount pipeline
  • Organizing data using the medallion architecture
  • Generating SQL transformations using natural-language interactions
  • BigQuery DataFrames and notebook

Hands-on Lab:

  • Generate SQL transformations.

Demo:

  • Medallion architecture, BigQuery DataFrames, and notebooks.

  • Pipeline orchestration with the Data Agent Kit
  • Building DAGs with natural-language instructions
  • Scheduling pipeline execution and GitHub Actions deployment
  • Monitoring pipeline execution through the agent

Hands-on Lab:

  • Build a DAG using natural language.

Demo:

  • Scheduling, GitHub Actions deployment, and monitoring walkthrough

  • Investigating pipeline incidents with the data agent
  • Tracing lineage to resolve reconciliation gaps
  • Diagnosing pipeline failures
  • Relationships between pipeline execution, lineage, and data issues

Hands-on Lab:

  • Investigate reconciliation gaps.

Scenario:

  • Resolve a pipeline incident and diagnose pipeline failures.

  • Modeling discount elasticity with BigQuery ML
  • Training the discount elasticity model within the data engineering workflow
  • Inference using the trained model
  • Model drift monitoring

Guided activity:

  • Train the discount elasticity model

Demo:

  • Model inference and drift monitoring.

  • Securing environments for data agents
  • Applying least privilege with the MCP Tool User role
  • Principal access boundaries and IAM deny policies for MCP
  • The prompt injection threat model and guarding against prompt injection
  • Document guardrails and securing data used by agent workflows
  • Configuring Model Armor

Hands-on Lab:

  • Least-privilege configuration

Demo:

  • Principal access boundaries, IAM deny policies, document guardrails, and Model Armor

  • Course review and knowledge validation
  • Reinforcing agentic data engineering workflows, MCP connectivity, and data service integration
  • Reviewing pipeline orchestration, troubleshooting, security, and governance practices
  • Scenario-based knowledge check and final quiz

Select Course date

Loading Dates...
Add to Wishlist

Course ID: 31195

Course Price at

Loading price info...
Enroll Now

FAQs for Orchestrate Workflows with the Data Agent Kit

This course is designed for intermediate-to-advanced practitioners working with data engineering, cloud data platforms, AI/ML, or agentic application development.

The Data Agent Kit is an agentic data engineering capability built on the Agent Development Kit (ADK) that connects AI-driven workflows with Google Cloud data services.

The course covers integrations with BigQuery, Spanner, BigLake, Dataproc, Managed Airflow, and Cloud SQL.

Yes. MCP is a key component of the course. Learners configure authentication and MCP connectivity and work with both local and remote MCP servers.

Yes. The course introduces ADK patterns and includes building a basic ADK data agent.

Yes. The course covers catalog browsing, universal search, and chat-driven schema, lineage, and data quality inspection.

Yes. Learners generate transformations, organize data using the medallion architecture, generate SQL transformations, and work with BigQuery DataFrames and notebooks.

Yes. Learners orchestrate pipelines, build DAGs using natural language, schedule workflows, deploy using GitHub Actions, and monitor pipelines through the agent.

Yes. The course covers resolving pipeline incidents, tracing lineage to identify reconciliation gaps, and diagnosing pipeline failures.

Yes. Learners work with BigQuery ML to model and train a discount elasticity model and explore inference and drift monitoring.

Yes. Security topics include secure agent environments, least privilege, the MCP Tool User role, principal access boundaries, IAM deny policies for MCP, prompt injection protection, document guardrails, and Model Armor.

The source course is listed as 5 hours 45 minutes and is classified as Advanced.

Yes. The course includes environment setup, ADK data agent development, MCP connectivity, data exploration, transformations, pipeline orchestration, troubleshooting, machine learning, and security activities.

Organizations can accelerate data engineering workflows through natural-language interactions, reduce manual pipeline and transformation development effort, improve data discovery and troubleshooting, automate orchestration workflows, integrate AI into existing data platforms, and establish stronger security controls for enterprise data agents.

Talk to Us