Hindi Voice Data Annotator

Audio Transcript Review & Correction — Hindi and English We are an AI company building high-quality speech datasets. We need native Hindi speakers to review and correct machine-generated transcripts of audio recordings. THE WORK You receive an audio file and a machine-generated (ASR) transcript of it. You listen, and you fix the transcript so it exactly matches what was said —misheard words, punctuation, speaker labels, numbers, names, and terms the machine got wrong. You are editing an existing draft, not typing from scratch. We provide the tool, the guidelines, and a worked example set. No prior transcription software experience needed. SCOPE & TIMELINE: - Pilot batch: 3 hours of audio, starting immediately - We are hiring for an initial pilot with strong potential for ongoing work at significantly higher volume for contractors who perform well - Fully remote, flexible hours, no fixed schedule - Work is assigned in batches; you claim what you can complete REQUIREMENTS: - Native or fully bilingual fluency in Hindi, including regional accents and colloquial speech - Strong written Hindi — correct spelling, punctuation, and grammar - Professional working and written English (our guidelines and team channel are in English) - Headphones and a quiet environment - Reliable computer and internet — this is browser-based work - Careful, consistent attention to detail over long sessions. This work rewards accuracy, not speed. NICE TO HAVE: - Prior transcription, subtitling, captioning, translation, or proofreading work - Experience with linguistics, journalism, or court/medical reporting HOW WE WORK - All communication happens in a shared Slack channel. You will be invited on day one. We ask that you check it once per working day. - We respond to questions within a few hours during US business hours. - You get written guidelines plus annotated examples before your first paid batch. Ask questions early — we would much rather answer them than fix work. - All work is covered by a standard NDA, signed before your first batch. The audio is confidential. We move fast. Expect a decision within 48 hours of your test submission. TO APPLY: Please answer these three questions at the top of your proposal. Proposals that do not answer them will not be reviewed. 1. What is your native language, and what region or dialect are you from? Applications are reviewed in the order received.