Google Cloud x DeepMind: What's New for Startups? #2

​In this second edition of our deep-dive series, we are moving beyond the chatbox to look at the underlying primitives of the Gemini 3 ecosystem. Join Thor Schaeff and Ian Ballantyne from Google DeepMind for a technical session focused on the latest model releases, the infrastructure enabling autonomous agents, and the exciting launches / announcements from Cloud Next!

​The Format:

• ​20 Minutes: Technical presentation & live architectural demos.

• ​25 Minutes: Open Q&A (Seed your questions in the registration form).

​​The "Latest from Lab" Agenda

• ​Gemini 3.1 Flash TTS (Released April 15): A technical look at the new gemini-3.1-flash-tts-preview. We’ll skip the hype and show you how to implement granular audio tags (e.g., [whispered], [shouting], [excited]) to programmatically direct vocal performance and multi-speaker dialogue.

• ​Gemma 4 for Local Agents: Our guest engineer will break down the Gemma 4 31B and 26B A4B architectures. We’ll discuss how its native support for tool-calling and complex reasoning makes it the new standard for local-first agentic intelligence.

​​Featured Speakers

• ​Thorsten Schaeff, Google DeepMind: Focusing on frontier research implementation.

• ​Ian Ballantyne, Google DeepMind Direct from the Gemma DevX team to discuss open-model optimization.

​​Access Note: No marketing fluff. We assume a baseline understanding of multimodal inputs/outputs. This is for technical founders and CTOs building at the frontier.