Tutorial: Google ADK & Cloud Run: AI Agents at Scale | Future of Data and AI | Agentic AI Conference

Why It Matters

It shows how serverless, agentic AI can be built at scale with minimal cost, enabling businesses to automate content creation and accelerate AI product deployment.

Key Takeaways

• Set up free $5 Google Cloud credit before starting the lab.
• Deploy each AI agent to its own Cloud Run instance for scaling.
• Use loop‑agent pattern: researcher ↔ judge before handing off to writer.
• Agents communicate via A2A HTTP protocol; orchestrator maintains context.
• Leverage Vertex AI Gemini and UV manager for rapid dependency installation.

Summary

The session walks participants through building a distributed, multi‑agent AI system on Google Cloud using the ADK and Cloud Run. Attendees first claim a $5 credit, create a new project, attach billing, and enable required APIs before cloning a starter repository and installing dependencies with the ultra‑fast UV package manager. Key steps include configuring environment variables, deploying each agent—researcher, judge, writer, and orchestrator—as separate Cloud Run services, and wiring them together via the A2A HTTP protocol. The loop‑agent pattern ensures the researcher and judge iterate until the output meets quality standards before the writer drafts the final course content. The tutorial highlights concrete code snippets: the researcher leverages a built‑in Google Search tool, the judge applies a structured schema to evaluate results, and the orchestrator preserves context across calls. Participants see real‑time debugging tips, such as resetting the project ID and verifying billing association. By modularizing agents and using serverless scaling, developers can rapidly prototype high‑quality, AI‑generated educational content that can be production‑ready, secure, and cost‑effective for enterprises seeking to automate knowledge creation.

Original Description

This hands-on lab by Shir Meir Lador, AI Leader at Google, shows how to move from single-agent setups to fully distributed systems. You will design and deploy a scalable multi-agent architecture using Google ADK, the A2A protocol, and Cloud Run, simulating real production workflows.

In this session, you’ll learn to:

- Build and orchestrate specialized agents using Google ADK and the Agent-to-Agent (A2A) protocol to enable structured, autonomous feedback loops

- Deploy a fully functional, scalable multi-agent system on Google Cloud Run for production-ready distributed architectures

_____

Learn data science, AI, and machine learning through our hands-on training programs: https://www.youtube.com/@Datasciencedojo/courses

Check our community webinars in this playlist: https://www.youtube.com/playlist?list=PL8eNk_zTBST-EBv2LDSW9Wx_V4Gy5OPFT

Check our latest Future of Data and AI Conference: https://www.youtube.com/playlist?list=PL8eNk_zTBST9Wkc6-bczfbClBbSKnT2nI

Subscribe to our newsletter for data science content & infographics: https://datasciencedojo.com/newsletter/

Love podcasts? Check out our Future of Data and AI Podcast with industry-expert guests: https://www.youtube.com/playlist?list=PL8eNk_zTBST_jMlmiokwBVfS_BqbAt0z2