When does class start/end?
Class hours may vary, please reach out to contact@ascendientlearning.com if you have any questions.
Building AI Agents with Google ADK
Build durable AI agents with Google ADK through workflow agents and Gemini Enterprise Agent Platform deployment. Agent anatomy and structured outputs cover ADK runtime primitives, model backends...
Read More
Duration
3 days
Course Code
GAI-2110
Available Formats
Classroom
Overview
Course Description
Build durable AI agents with Google ADK through workflow agents and Gemini Enterprise Agent Platform deployment. Agent anatomy and structured outputs cover ADK runtime primitives, model backends through LiteLlm, Pydantic schemas and validators, and the choice between schema-constrained and tool-using output. Tools, sessions, and memory cover function and agent-as-tool registration, runtime context, MCP integration, swappable session backends, managed memory backends with Agent Platform Memory Bank, streaming, and human-in-the-loop confirmation. Evaluation, orchestration, and deployment cover the callbacks and plugins extension model, evaluation sets and LLM-as-judge metrics, sequential, parallel, and loop workflow agents, multi-agent coordination through A2A, and Agent Runtime through the adk CLI. Hands-on labs produce a multi-tool ADK agent, an evaluation pipeline with a guardrail plugin, a multi-agent A2A workflow, and a Agent Runtime deployment. The course is designed for software developers, AI engineers, and data scientists with Python experience and a Google Cloud account.
Skills Gained
By the end of this course, participants will be able to:
- Build ADK agents with structured outputs, tools, and runtime context
- Manage multi-turn conversations through managed memory, streaming, plus human-in-the-loop patterns
- Evaluate ADK agent quality through callbacks, plugins, and evaluation sets
- Orchestrate ADK workflow agents plus multi-agent A2A systems with error recovery
- Deploy ADK agents to Agent Runtime with observability tracing
Who Can Benefit
This course is designed for:
- Software Developers
- Data Scientists
- AI Engineers
Prerequisites
Participants should enter this course with:
- Practical Python experience
- GAI-1101 or equivalent
- Google Cloud account with billing enabled
Organizational Objectives
This course assists organizations to:
- Lower agent error rates through Pydantic schemas, callback guardrails, and evaluation sets
- Reduce time-to-production for ADK agents through Agent Runtime deployment
- Build reusable agent patterns across projects through workflow agents and sub-agents
- Establish a shared discipline for agent observability through OpenTelemetry tracing
Software
All attendees must have a modern web browser and an Internet connection.
Course Details
Course Details
Module 1 - Anatomy of an ADK Agent
By the end of this module, you will be able to create and run a basic ADK agent across multiple model backends, configure static, dynamic, and state-templated instructions, and inspect agent events and traces in the ADK Dev UI.
- The ADK runtime model — agents, runners, events, and sessions
- Creating and running a basic agent
- Instructions — static, dynamic, and state-templated
- Configuring model backends — Gemini and external models via LiteLlm
- Synchronous and asynchronous execution patterns
- The adk CLI and Dev UI for event and trace inspection
- Hands-on Lab: Build a first ADK agent across multiple model backends, run it synchronously and asynchronously, and trace its events in the Dev UI.
Module 2 - Structured Outputs with ADK
By the end of this module, you will be able to constrain ADK agent responses with Pydantic input and output schemas, handle validation failures and retries, and capture structured results with named result keys for downstream consumers.
- Defining input and output schemas with Pydantic
- Constraining responses with schemas and validators
- Handling validation failures and retries
- Schema-constrained output versus tool-using output as a design choice
- Multi-format responses with union schemas and specialized agents
- Capturing structured results with named result keys
- Hands-on Lab: Build an ADK document classifier that produces validated structured outputs from unstructured input.
Module 3 - Tools and External Integrations with ADK
By the end of this module, you will be able to register tools as functions or nested agents, pass runtime context through the tool context object, and connect external tools through the Model Context Protocol.
- The tool lifecycle — definition, validation, and execution
- Tool registration — functions and nested agents
- Passing runtime context through the tool context object
- Tool error handling and graceful degradation
- Connecting external tools via the Model Context Protocol
- Hands-on Lab: Build a multi-tool ADK agent that queries BigQuery and invokes a custom Python function tool.
Module 4 - Sessions, State, and the Runtime Context
By the end of this module, you will be able to use sessions, events, and state as ADK’s runtime substrate, scope state across tools and sub-agents, and swap session backends from local in-memory to managed Agent Platform Sessions or Firestore.
- Sessions, events, and state as the runtime substrate
- State scoping by lifetime — temporary, app, user, and session-scoped
- Sharing state across tools and sub-agents
- Session backends — in-memory, database, Agent Platform Sessions, and Firestore
- Artifact service for files and binary data
- Hands-on Lab: Build a configurable ADK agent with swappable session backends and shared state across tools.
Module 5 - Conversations, Memory, and Streaming
By the end of this module, you will be able to manage multi-turn conversations with message history, persist long-term knowledge through managed memory backends including Agent Platform Memory Bank, and stream responses with human-in-the-loop confirmation patterns.
- Managing multi-turn conversations with message history
- Long-term knowledge with managed memory backends and Agent Platform Memory Bank
- Streaming responses — token-level and structured
- Human-in-the-loop confirmation patterns
- Distinguishing short-term session state from long-term memory
- Hands-on Lab: Build a streaming conversational ADK agent with persistent memory across Agent Platform Memory Bank.
Module 6 - Callbacks, Plugins, and Evaluation
By the end of this module, you will be able to attach agent-scoped callbacks and process-wide plugins as guardrails, run automated evaluation against evaluation sets, and detect hallucination and drift with LLM-as-judge metrics.
- Defining success criteria for agent outputs
- Callback lifecycle hooks at agent, model, and tool boundaries
- Plugins — process-wide behavior for logging, caching, and reflect-and-retry
- Input and output guardrails through callbacks and plugins
- Automated evaluation with evaluation sets, trajectory, and response criteria
- Hallucination and drift detection with LLM-as-judge metrics
- Hands-on Lab: Build an ADK evaluation pipeline with a guardrail callback at the model boundary and a process-wide plugin for logging.
Module 7 - Workflow Orchestration with ADK
By the end of this module, you will be able to compose deterministic workflows with ADK’s sequential, parallel, and loop workflow agents, recover from errors with a reflect-and-retry pattern, and resume long-running runs through checkpointing and rewind.
- Sequential, parallel, and conditional patterns
- Workflow agents — sequential, parallel, and loop
- Error propagation, recovery, and reflect-and-retry strategies
- Checkpointing, session resumption, and rewind
- Choosing deterministic workflows over LLM-driven coordination
- Hands-on Lab: Build a research-and-report workflow with ADK workflow agents, branching, and reflect-and-retry error recovery.
Module 8 - Multi-Agent Systems and A2A
By the end of this module, you will be able to recognize when to split a problem into multiple ADK agents, build hierarchical teams with sub-agents, and expose or consume remote agents through the A2A protocol.
- When to use multiple agents versus a single agent
- Plugins in a multi-agent context
- Hierarchical agent teams with sub-agents
- Shared versus isolated context across an agent hierarchy
- LLM-driven agent transfer versus explicit agent-as-tool invocation
- Exposing and consuming remote agents via the A2A protocol
- Hands-on Lab: Build a multi-agent ADK workflow with delegation and handoff between local and remote agents.
Module 9 - Testing and Deploying ADK Agents
By the end of this module, you will be able to unit-test ADK agents with mocked model backends, capture observability through OpenTelemetry spans exported to Cloud Trace, and deploy agents to Agent Runtime with the adk CLI.
- Unit-testing agents with mocked model backends and pytest assertions
- Observability via OpenTelemetry spans exported to Cloud Trace
- Serving agents as FastAPI endpoints from the adk CLI
- Deploying to Agent Runtime with the adk CLI
- Usage tracking and cost-quality trade-offs in production
- Hands-on Lab: Unit-test an ADK agent with mocked models, then deploy it to Agent Runtime.
Schedule
FAQ
Does the course schedule include a Lunchbreak?
Classes typically include a 1-hour lunch break around midday. However, the exact break times and duration can vary depending on the specific class. Your instructor will provide detailed information at the start of the course.
What languages are used to deliver training?
Most courses are conducted in English, unless otherwise specified. Some courses will have the word "FRENCH" marked in red beside the scheduled date(s) indicating the language of instruction.
What does GTR stand for?
GTR stands for Guaranteed to Run; if you see a course with this status, it means this event is confirmed to run. View our GTR page to see our full list of Guaranteed to Run courses.
How do I find an Ascendient Learning training location?
We have training locations across the United States and Canada - View a complete list of classroom training locations.
Which delivery formats are available?
At Ascendient Learning, we offer training that is Instructor-Led, Online, Virtual, and Self-Paced.
Does Ascendient Learning deliver group training?
Yes, we provide training for groups, individuals and private on sites. View our group training page for more information.
What does vendor-authorized training mean?
As a vendor-authorized training partner, we offer a curriculum that our partners have vetted. We use the same course materials and facilitate the same labs as our vendor-delivered training. These courses are considered the gold standard and, as such, are priced accordingly.
Is the training too basic, or will you go deep into technology?
It depends on your requirements, your role in your company, and your depth of knowledge. The good news about many of our learning paths, you can start from the fundamentals to highly specialized training.
How up-to-date are your courses and support materials?
We continuously work with our vendors to evaluate and refresh course material to reflect the latest training courses and best practices.
Are your instructors seasoned trainers who have deep knowledge of the training topic?
Ascendient Learning instructors have an average of 27 years of practical IT experience and have also served as consultants for an average of 15 years. To stay current, instructors spend at least 25 percent of their time learning new, emerging technologies and courses.
Do you provide hands-on training and exercises in an actual lab environment?
Lab access is dependent on the vendor and the type of training you sign up for. However, many of our top vendors will provide lab access to students to test and practice. The course description will specify lab access.
Will you customize the training for our company’s specific needs and goals?
We will work with you to identify training needs and areas of growth. We offer a variety of training methods, such as private group training, on-site of your choice, and virtually. We provide courses and certifications that are aligned with your business goals.
How do I get started with certification?
Getting started on a certification pathway depends on your goals and the vendor you choose to get certified in. Many vendors offer entry-level IT certification to advanced IT certification that can boost your career. To get access to certification vouchers and discounts, please contact info@ascendientlearning.com.
Will I get access to content after I complete a course?
You will get access to the PDF of course books and guides, but access to the recording and slides will depend on the vendor and type of training you receive.
How do I request a W9 for Ascendient Learning?
View our filing status and how to request a W9.
Reviews
Class was easy to sign up for and ExitCertified provided very good communication
I registered a day before class and am happy that I received all the materials and links in time for the class. Thanks.
The instructor really took his time and made sure I was able to understand the concepts.
The exit certified aws course provided a good introduction to the tools available on aws.
great class and packed with material, would have lived to spread it more into many days but overall very informative.
