What the work involves
You receive Telugu audio or video clips — often spontaneous speech, phone-quality recordings, code-switched Telugu-English, or regional dialect samples — and produce written transcripts that follow a client style guide down to the detail: how to render disfluencies, whether to transliterate English loanwords or keep Latin script, how to timestamp, when to tag a segment as inaudible versus guess. Alongside raw transcription you'll segment and label audio, review and correct transcripts other annotators produced, and file notes when the guideline doesn't cover what you actually heard. A meaningful share of the job is the last part: flagging ambiguity rather than silently resolving it.
What the platform screens for
micro1's screening is AI-led and conversational. Expect it to test Telugu depth directly — dialect recognition (Telangana vs. coastal Andhra vs. Rayalaseema), handling of Sanskritised versus colloquial registers, orthographic decisions around anusvāra, gemination, and English insertions. It will also probe judgment: what you do when two speakers overlap, when a speaker uses a word you don't know, when the guideline and natural Telugu conflict. Prior AI experience is explicitly not required, but demonstrated transcription or linguistic-data work is weighted heavily, and vague claims of fluency without examples tend not to survive follow-up.
Logistics
- Fully remote, contractor engagement, invoiced hourly or per task depending on the batch.
- Largely asynchronous — you pull work from a queue against milestone deadlines rather than sitting in scheduled shifts.
- Volume fluctuates with client demand; treat it as supplemental rather than guaranteed load.
- You'll need reliable internet, decent headphones, and comfort with a browser-based annotation tool and a Telugu keyboard input method.
- Pay in the $10–24/hr band reflects what contributors have reported; rate varies by task complexity, review versus first-pass work, and quality scores.