Building Conversational Agents — Thor Schaeff and Philipp Schmid, Google DeepMind

By ai.engineer

Categories: AI, Tools

Summary

Google DeepMind's developers showcase how to build conversational agents with Gemini API at no cost—the free tier covers all hands-on demos in this workshop. Learn practical setup, API key management, and real-world deployment patterns from engineers shipping production AI features.

Key Takeaways

  1. Access Gemini API free tier with just a Gmail account—no credit card required. Visit ai.dev or aistudio.com to create API keys in minutes for hands-on development.
  2. Treat API keys as secrets like passwords. Developers frequently leak keys to GitHub through committed code—use environment variables (.bashrc, .zshrc) or inline injection instead.
  3. Google AI Studio provides a zero-code interface to test Gemini models before building API integrations. Ideal for rapid prototyping before committing engineering resources.
  4. Gemini supports WebSocket connections for real-time conversational interactions—enabling live voice and multi-language agent capabilities demonstrated via phone software.
  5. Multi-language support is native to Gemini. Workshop attendees from 10+ countries tested models across Spanish, Romanian, Czech, Farsi, and Hindi in real-time.

Topics

Transcript Excerpt

Hello everyone. >> Hello. >> Perfect. >> It's just that Philip and I were both Germans and we thought it was funny. Maybe we we can do it in German. Actually, looks like there's a there's a German crew there, which is nice. Uh, no, no worries. We'll we'll we'll do it in English. We'll do it in a couple different languages. Maybe we'll find out. Do we have other languages in the room? >> Other nationalities? Yeah. What do we have? >> Shout it out. >> So, Spanish. >> Any Icelandic? >> No. >> D. Ok...