What the work actually involves

You will spend your day producing and evaluating analytical reasoning tasks — the kind of problem a model gets almost right and then fumbles. That means reading a sales distribution across months and locations and reasoning out which site genuinely grew most (and defending your choice of time window), or working a constrained-scheduling puzzle where visit rules interact. For each item you supply the correct answer plus an explanation clear enough that a model can learn the reasoning path, not just the result. A second stream of work is critique: reading model output, identifying exactly where the chain of reasoning breaks, and writing annotations a trainer can act on. Because the role is specified as English and Bahasa Indonesia, expect tasks that live in both languages — writing or reviewing Indonesian prompts and responses, and checking that meaning, numeric detail and register survive between the two.

What the screen is looking for

Turing states no specialized domain background is required, which shifts the whole weight of the screen onto demonstrable reasoning and writing. Expect to be asked to break a long piece of content into logical blocks, verify a factual claim through online research, and show basic quantitative comfort — percentages, growth rates, reading a table without misreading it. Genuine working fluency in Bahasa Indonesia is a gate, not a nice-to-have, and it is usually tested by asking you to produce or judge Indonesian text rather than by asking whether you speak it. Prior professional writing work — business or research analyst, journalist, technical writer, editor, translator — is listed as preferred and is the most common way candidates evidence the explanation-writing half of the job.

Logistics and honest expectations

  • Contractor engagement, no medical or paid leave, maximum 12 months with extension tied to performance and project need.
  • 40 hours per week, fully remote, with 2–5 hours daily overlapping America/Los_Angeles (UTC-8).
  • Own desktop or laptop and a reliable connection; Excel and Google Suite familiarity is expected.
  • Pay is undisclosed on this listing. Turing describes compensation as based on experience and expertise; treat any figure you see quoted elsewhere as observed, not promised, and confirm the rate in writing before you commit.

The realistic upside is exposure to how frontier labs actually build reasoning data, and a habit of writing explanations precise enough to survive an adversarial reader. The realistic downside is volume: this is a high-throughput annotation job, and the tasks repeat.