What the work involves
This is a single, self-contained recording task rather than an ongoing engagement. You receive a short script and a written guide covering room noise, mic distance, file format and pacing, read the script aloud in your own voice, and upload the file. Most contributors finish in under 15 minutes end to end. Occasionally a submission is returned for a re-record — usually because of background noise, clipping, or a skipped line — and re-recording is part of the task.
The project collects authentic Slovak-influenced English speech for speech-model evaluation and training. That means the accent has to be yours: someone who grew up speaking Slovak and learned English, or who speaks English daily with a Slovak substrate. Performed or approximated accents are the specific failure mode this project screens hardest against, because they poison the dataset they are meant to improve.
What the screen looks for
- Verifiable Slovak language background — where you learned Slovak, where you learned English, what you speak day to day
- Working audio capture: a phone or laptop mic in a quiet room is enough; no studio required
- Ability to read English instructions and act on them without back-and-forth
- Willingness to re-record once if the first file has a technical fault
The screen is short and AI-led. It will probe your language history more than your recording gear, and it will ask you to describe your actual setup in concrete terms rather than confirm that you have one.
Logistics
Fully remote and fully asynchronous — there is no call, no scheduled session, and no fixed hours. Turnaround is quick: contributors are generally asked to submit within a day or two of receiving the script. Pay is a flat $50 for a single accepted submission; whether further recordings are offered depends on the project, so treat this as a one-off unless AfterQuery says otherwise.