Vibe Coding and Context Engineering with Agent Development Kit

Learn to build and deploy AI agents from scratch using natural language through Google's innovative "vibe coding" technique with the Gemini CLI and Agent Development Kit (ADK). Follow along as instructors Sita and Amit demonstrate how to scaffold a complete expense tracker AI agent using only plain English commands, then master "Context Engineering" - the critical methodology for transforming prototypes into production-ready applications. Discover how to provide documentation, coding standards, and feature requirements to guide AI in generating high-quality, maintainable code, then witness a seamless one-command deployment to Google Cloud Run. Explore advanced topics including refactoring code into tools and models, debugging Python applications with Gemini CLI, creating essential context files like Gemini.md and PRDs, and preparing applications for cloud deployment with Docker. The session concludes with an "Eng Talk" segment addressing key community questions about RAG relevance with million-token context windows, managing context overload, tracking context retrieval metrics, and optimizing token usage for maximum efficiency.