What the work involves

AfterQuery is collecting authentic accented English speech to train and evaluate speech models. Your contribution is a single recording session: you receive a short English script and a written recording guide, read the script aloud in your own natural Kannada-influenced English, and upload the audio file through the platform. Start to finish, most contributors finish in under 15 minutes.

  • Read a provided script in English, aloud, at a natural conversational pace
  • Follow the guide on microphone distance, room noise, file format and length
  • Re-record if the guide's checks (background noise, clipping, truncation) aren't met
  • Submit through the designated upload link, not by email or messaging app

What the platform screens for

The screen is short and mostly about authenticity and technical viability. Reviewers want speakers whose everyday English genuinely carries Kannada influence — typically people who grew up speaking Kannada or lived long-term in Karnataka. Imitated or performed accents are rejected, and this is checked by ear at review. The other filter is practical: a working microphone, a quiet enough space, and the ability to read written English instructions and act on them without back-and-forth support.

Logistics

  • Fully remote and asynchronous; record whenever you like within the turnaround window
  • One-off task, not ongoing work — flat $50 per accepted recording, as observed on the listing
  • No studio equipment required; a modern smartphone in a quiet room is sufficient
  • Turnaround is quick, so expect to complete and submit within a day or two of being assigned