Hands-On Graph Based Workflows with ADK 2.0: From Agentic AI to Agentic Engineering

Unlock the potential of your AI agents by transitioning from simple prompt chains to robust, graph-based workflows. In this video, we dive deep into architectural patterns that enable deterministic control, branching logic, and loop management in agent systems.

Key concepts covered:

* Defining Agent Workflows: Combining AI reasoning with deterministic code logic.
* Graph Architecture: Structuring agent logic as a graph for complex, multi-agent systems.
* Execution Nodes: Utilizing Agents, ADK Tools, and Custom Functions as processing units.
* Event-Driven Communication: Passing data through inputs, outputs, messages, and persistent state across your workflow graph.
* Input & Output Schemas: Ensuring production-grade reliability by constraining data formats with Pydantic-style models.

Learn how to evolve your AI development from basic prompts to production-ready, structured applications.

Resources:
GitHub Repo: https://github.com/10xroadmap/google-adk-2.0-examples
Follow for more: https://youtube.com/@10XROADMAP

Видео Hands-On Graph Based Workflows with ADK 2.0: From Agentic AI to Agentic Engineering канала 10XROADMAP