Submission Checklist

Required by the hackathon. Check every item from a clean session before submitting.

Repository

• Public GitHub repo with source and documentation
• README: prerequisites, dependencies, exact setup/run steps, dataset access, known limitations
• README discloses AI coding tools and which parts were AI-assisted
• A clean clone installs, accesses the dataset and runs the demo from the README

Dataset

• Audio, exact transcripts, provenance, reviewed alignment and flaw labels in GitHub or a public Google Drive folder linked in the README
• Reuse rights checked for each source recording
• Review status and tolerance stated honestly

Application

• Dashboard demonstrates audio and transcript upload, processing, time-series comparison, and text feedback
• Errors and low-confidence states are shown, never hidden

Technical document (max 6 pages)

• Dataset construction, evaluation rubric, architecture, scoring, evaluation results, limitations
• Every metric lists sample count and label tolerance

Demo video

• 3–10 minutes, shows the project running, explains approach and stress tests
• English audio or subtitles
• Visibility Public or Unlisted (not Private)

Devpost

• Description covers problem, solution, build
• Built With lists AI coding tools
• Every teammate has an account and is added with their real full name

Final verification

• Repository, dataset, app instructions and video links opened from a clean session
• Submitted before the team cutoff on Oct 14