Workshop Structure

Part 1 - Build: Basic voice pipeline

• Set up STT → LLM → TTS flow
• Create simple "Hello World" voice agent
• Test with basic commands and responses

Key Learning: Understanding the core architecture and data flow

Part 2: Intelligence

• Integrate sophisticated LLM reasoning
• Add conversation memory and context management
• Handle multi-turn conversations

Key Learning: Making agents conversational vs. transactional

Part 3: Production Polish

Build: Robust, deployable system

• Implement error handling and graceful degradation
• Add retry logic and fallback strategies
• Component selection and optimization techniques

Key Learning: Production-ready code vs. demo code

Part 4: Evaluation & Deployment

Build: Monitoring and evaluation framework

• Set up performance metrics and logging