Level 6: Applied: cases, FDE scenarios, mocks · L6.3 AI-era cases
Real-time voice agent
Budget latency across speech-to-text, LLM and text-to-speech.
15 min · Pro
What you take away
- The whole budget is about 800 ms from the caller finishing to the first audio back, and the four stages sum to 800–950 ms with no slack anywhere.
And 3 more inside the class, in Pro.
How the class is laid out
- The prompt
- Requirements and the numbers
- The design
- Three decisions
- The twist
- Check your understanding