I Built an AI Dietician That Lives on WhatsApp 🥗 — Try It Yourself! (Google ADK + Gemini + K8S)

This project demonstrates building a live AI dietician agent deployed on WhatsApp using Google's Agent Development Kit (ADK), Gemini API, and Kubernetes. The agent provides real-time nutritional guidance and dietary recommendations through WhatsApp's messaging interface. The implementation showcases end-to-end deployment of an AI agent from development to production on K8S infrastructure.

Key Points

• •Use Google ADK (Agent Development Kit) to scaffold and manage AI agent development with built-in integrations
• •Leverage Gemini API as the LLM backbone for nutritional knowledge and personalized dietary recommendations
• •Integrate WhatsApp as the user-facing interface for seamless, no-signup access to the AI dietician
• •Deploy the agent on Kubernetes (K8S) for scalable, production-ready infrastructure management
• •Implement message handling and state management to maintain conversation context across WhatsApp interactions
• •Use Google Cloud services (likely Cloud Run or GKE) for hosting and orchestrating the agent workloads
• •Design the agent with domain-specific prompting for accurate nutritional and dietary guidance
• •Enable real-time interaction without requiring users to install apps or create accounts

Workflow Diagram

Concepts