What the work involves

This is a single, self-contained recording task, not an ongoing engagement. AfterQuery supplies a short English script and a recording guide covering things like room noise, mic distance, pacing, and file format. You read the script aloud in your own voice, check the playback against the guide, and upload the file through the designated submission link. Most contributors finish in 10–15 minutes, including the read-through.

The data is used to train and evaluate speech systems on accented English — a category where models routinely underperform. That is why the requirement is a genuine Russian-influenced English accent, not a performed one: imitated accents introduce artefacts that make the recording useless for the purpose, and they are usually audible to reviewers.

What the platform screens for

  • Authenticity of accent. Expect to be asked about your first language, where you learned English, and how long you have spoken it. Some screens include a short live or recorded speech sample.
  • Recording setup. A smartphone is fine. A quiet room matters more than expensive equipment — expect questions about background noise, echo, and whether you can control your environment for fifteen minutes.
  • Instruction-following in written English. The guide is in English and there is no hand-holding; screens probe whether you can read a spec and comply with it without asking clarifying questions on every point.
  • Turnaround. These projects are opened with short deadlines. Being able to submit within a day or two of acceptance is often the deciding factor.

Logistics and pay

Fully remote and asynchronous — record whenever you like, from wherever you are, so long as the environment is quiet. Pay is observed at a flat $50 per accepted single-recording submission, which annualises to roughly $30/hr against the task's expected length; it is stated as flat per submission, not hourly, and re-records for quality issues are typically expected within that flat fee. There is no guarantee of follow-on work, though contributors who submit clean audio on time are frequently invited back for similar batches.