Fix Flutter iOS Voice Recognition

We have an existing voice recognition feature that is currently buggy. Your job is to clone our GitHub repo, reproduce the issue, identify the root cause, and ship a clean fix. This is a bug-fix + stabilization task (not a greenfield build).

What you’ll do
- Pull our GitHub repo (shared after hire) and set up the project locally.
- Reproduce the voice recognition bug on real device(s) and/or simulator.
- Debug and isolate the root cause (permissions, audio session, plugin/native bridge, streaming, model/pipeline, threading, etc.).
- Implement a fix that is stable and doesn’t break existing features.
- Add minimal tests/logging so we can verify and avoid regressions.
- Provide a short handover note: what was broken, why, and how you fixed it.

Expected Deliverables
- A working PR/branch with the fix.
- Clear steps to reproduce + verify the fix.
- Notes on root cause and any recommended follow-up improvements.

Nice to have
- Experience with Flutter speech/voice plugins or native iOS audio/speech APIs.
- Experience tuning or troubleshooting NLP / speech pipelines.
- Ability to add lightweight diagnostics (logs/metrics) to prevent silent failures.

How to apply

Please include:
1. A short explanation of how you would approach debugging this kind of issue.
2. One relevant project you’ve worked on (voice, audio, speech recognition, NLP, Flutter+iOS).
3. Your estimated timeline after you receive the repo and reproduction steps.