Week 6 Worklog

Work Completed

• Studied Amazon Transcribe job flow for audio-to-text processing.
• Reviewed audio input requirements, supported formats, language codes, and output locations.
• Added transcript upload fallback for cases where audio transcription is not reliable.

Results Achieved

• Added Transcribe to the architecture as an optional path for audio input.
• Defined en-US and vi-VN as initial language choices for testing.
• Documented that demo audio should be short, clear, and consent-based.

Learning Outcomes

• Speech-to-text quality depends on audio quality, noise, language, and speaking style.
• A good workshop should include fallback steps so another learner can still complete it.