The AI agent hype is deafening. The actual understanding of what an agent is: near silence.

Most developers have built a chatbot and called it an agent. They are not the same thing, and that gap matters. An agent perceives, decides, acts, and loops back. Once you see that pattern, you cannot unsee it, and you can build one in under an hour.

Featured as a core track of Smart City Innovation Day, this hands-on workshop cuts through the noise. Led by Tamas Piros (AI Consultant & Google Developer Expert), we will build a real agent from scratch, live, using TypeScript, Google's Agent Development Kit (ADK), and Gemini.

Instead of overwhelming you with all twenty-one design patterns that exist, we will focus on layering the architecture that actually works for real-world and sensor-based data: starting with a single LlmAgent and one tool, moving to a SequentialAgent for chained reasoning, and ending with a ParallelAgent for fan-out work with an aggregator. Every architectural decision gets explained as we make it, connecting agentic AI directly to smart public administration and municipal decision-support systems.

What You Will Walk Away With:

• Working Code: A fully functioning agent running live on the Gemini ecosystem.

• Production Roadmap: A clear, practical path to deploying your finished agent on Cloud Run.

• Core Architectures: Mastery of sequential, parallel, and multi-agent workflows without the fluff.

• A Solid Mental Model: The skills required to apply agentic workflows to tangible data visualization and infrastructure challenges.

Who Should Attend?

This event is designed for Computer Science, Software Engineering, and Data Science students, alongside young developers and innovators who want to build real systems rather than just talk about AI.