Add the setup options and mic check; rename the voice engine folder
main·
Commit

- Setup: video or phone and the interviewer (Ananya or Aarav, each with an avatar and Kokoro voice) fixed on Continue; track, level, round, number of questions, answer length, mood and an optional name chosen while models load; settings reach the interview engine through the session
- Mic check: why it is needed, a level meter, what Whisper heard, and the interviewer's voice
- Rename src/engine to src/voice (VoiceHost, PhoneVoice, VideoVoice) so 'engine' only means the interview engine
- Drop the unused interviewer caption state from the voice store
- Dev-only setup preview at ?dev=setup
Entire-Checkpoint: 01M41T0WCXE1HQ4QZ69H86Z0NT
Checkpoints
Checkpoint 1