What the work involves

You receive Norwegian audio and video — interviews, spontaneous conversation, read speech, sometimes phone-quality or noisy recordings — and produce written transcripts that hold to a specific style guide. That guide, not your instincts, decides how you handle filler words, false starts, overlapping speakers, numbers, dialect forms, and code-switching into English. A second strand of the work is reviewing other contributors' transcripts and correcting them against the same rules, plus labelling speech segments (speaker turns, noise events, unintelligible spans) according to the project's tag set.

The judgment that matters most is knowing when not to commit. If two seconds of audio could plausibly be two different words, the expected behaviour is to mark it with the prescribed unclear tag and note your best interpretation — not to write the version that reads more smoothly. Norway's dialect spread makes this sharper than in many languages: you will meet speakers whose spoken forms sit well outside Bokmål orthography, and the guideline will tell you whether to normalise or preserve.

What the platform screens for

micro1 runs an AI-led interview before any project work. Expect it to test native-level Norwegian command in concrete terms — orthography, punctuation of spoken syntax, Bokmål/Nynorsk awareness, dialect comprehension — and to probe how you resolve conflicts between a written guideline and what you actually hear. Prior AI or machine-learning experience is not required and claiming it earns nothing; demonstrable transcription discipline does. Screens typically include follow-up questions, so answers that name specific tools, audio conditions, and decision rules survive better than general claims of being detail-oriented.

Logistics

  • Fully remote contractor engagement, invoiced hourly; pay observed at $20–35/hr and not guaranteed for any given batch
  • Largely asynchronous: tasks are queued, with occasional coordinator threads to resolve guideline ambiguity
  • Flexible hours, but projects run on milestones — sustained weekly availability matters more than when you work
  • You supply your own machine, reliable internet, and good headphones; confidential audio must stay on approved systems and never be pasted into public AI tools