Building Effective AI Agents for Discovery & Innovation

About the Session

We have moved beyond the era of chatbots. In 2026, the most productive researchers and engineers aren't just prompting AI—they are building agentic systems that plan, use technical tools, and execute complex workflows autonomously.

Join the Google Cloud Advanced Computing Community for a tactical, deep-dive session on how to create effective AI agents. We will move past the hype to look at the specific architectures—Plan, Tools, Memory, and Evaluation (PTME)—required to build agents that are reliable enough for high-stakes business and technical environments.

What We Will Cover:

• The Google Agent Stack: Learn how to use Gemini 3 Pro’s reasoning engine alongside the Agent Development Kit (ADK) and the Antigravity environment in AI Studio to rapidly prototype agents that understand code, math, and massive datasets.

• Beyond the Google Ecosystem: We will also show how to use other AI technologies, and how to make different technologies work together with MCP, A2A, and more.

• Engineering for Innovation & Productivity: See real-world examples of agents designed to perform literature synthesis, automate data cleaning across BigQuery, and orchestrate jobs and workflows

• Reliability & Guardrails: How to implement "human-in-the-loop" checkpoints to debug your agent’s reasoning and ensure it stays on track.

Whether you want to build personal productivity agents or large-scale agentic systems for your organization, this session provides the blueprint for building agents that actually work.

Speaker