What the work involves

You spend your day building and grading analytical problems for a large language model. That means reading long source material and breaking it into logical blocks, checking claims against online sources, constructing questions the model is likely to get wrong, and then writing the correct answer with an explanation clear enough that a model can learn the reasoning path — not just the result. Typical items look like growth-rate comparisons across a sales table where the choice of time window changes the answer, or constraint puzzles with ordering rules where the trap is an unstated assumption. Tagalog is required alongside English, so expect work that spans both languages: prompts, responses, and annotation written or reviewed in Tagalog, with English as the working language of the project.

What the platform screens for

Turing states no prior specialized domain experience is needed, which shifts the screen onto demonstrable reasoning and writing. Screeners probe whether you can explain why an answer is right in steps someone else could audit, whether you notice ambiguity in a question instead of guessing past it, and whether your written feedback is specific enough to act on. Expect to be asked to work through a small data-interpretation or logic problem live, and to be pushed on your reasoning even when your answer is correct. Professional writing background — analyst, journalist, editor, translator, technical writer — is preferred and comes up often. Genuine working Tagalog is verified rather than assumed; conversational familiarity is usually not enough for annotation work.

Logistics

  • Contractor engagement: no medical coverage, no paid leave.
  • 40 hours per week for the contract duration, with 2–5 hours per day overlapping America/Los_Angeles (UTC-8).
  • Fully remote; you supply your own desktop or laptop and a reliable connection.
  • Pay is undisclosed on this listing and described by Turing as based on experience and expertise. Contract extension is possible based on performance and project needs.

The day is largely solo and async, punctuated by written feedback cycles from reviewers. People who do well here treat their own explanations as the deliverable, not a footnote to the answer.