What the work actually involves

You receive Finnish audio — likely a mix of scripted reads, spontaneous speech, phone-quality recordings and multi-speaker conversation — and produce verbatim transcripts that conform to a project style guide. That guide will specify things most people get wrong on first pass: how to render filled pauses and false starts, whether to normalise numbers and abbreviations, how to tag speaker turns and overlaps, what to do with code-switched English, and how to mark inaudible spans. A second strand of the work is review: reading another contributor's transcript against the audio, deciding whether a difference is an error or an acceptable variant, and correcting without rewriting to personal taste.

Finnish specifics matter here more than in most languages. Spoken Finnish diverges sharply from the written standard, and projects differ on whether mä tuun is transcribed as heard or normalised to minä tulen. Dialectal features — Savo, Tampere, Turku, Ostrobothnian, Helsinki slang — plus sandhi, consonant gradation in casual speech, and heavy assimilation all produce judgement calls. You are expected to flag ambiguity rather than guess quietly, and to write up recurring problems so the guideline can be tightened.

What the platform screens for

micro1's process is AI-led: an asynchronous interview that asks about your background, then probes Finnish specifically with follow-up questions, and usually pairs with a short transcription or review sample. Screeners are looking for native-level command of both standard and colloquial Finnish, evidence you have worked to a written style guide before, and the instinct to escalate an ambiguity instead of inventing a plausible reading. No AI or machine-learning background is required.

Logistics

  • Fully remote, contractor engagement, invoiced hourly or per task depending on the batch.
  • Largely asynchronous with assignment deadlines and milestones; some written back-and-forth with project leads.
  • Volume is project-driven and fluctuates — treat it as flexible supplementary work rather than a guaranteed weekly load.
  • You will need reliable internet, a quiet workspace, and decent over-ear headphones; some batches require specific transcription tooling.
  • Pay of $20–35/hr reflects observed rates for this role and is not guaranteed for any given batch.