Audio Transcription Specialist

Audio Transcription & Alignment Specialist Expertious.ai · Hourly contract · Remote · 10–40 hrs/week About the role Speech models break on the things humans handle without thinking. Crosstalk. A swallowed consonant. The half-second where a false start turns into a real word. A model only learns where one word ends and the next begins if someone marked that boundary correctly in the first place. That's this job. You'll work inside a proprietary editor, correcting word-level transcripts and aligning timing boundaries to the millisecond — word by word, frame by frame. The work is exacting and it is directly upstream of how well these systems hear people. We're hiring in Sydney because the city's media, film, and language services industries turn out people with exactly this ear. If you've done closed captioning, broadcast transcription, subtitling, linguistics, or audio annotation, you already have the instincts. No AI background required. Fully remote. You set your own hours. What you'll do Correct word-level transcripts for accuracy, spelling, grammar, and strict style conventions Align word and segment boundaries to the spoken audio at millisecond precision Edit text and timing simultaneously inside a specialized web-based labeling tool Apply detailed style guides consistently across every task, without drift Resolve the hard cases: overlapping speech, filler words, false starts, background noise, unintelligible audio Hold quality across high task volume and tight per-task turnarounds Work independently and asynchronously Who you are Native or near-native English fluency with excellent listening comprehension Detail-obsessed at a level most people find excessive — you hear what others skip past Consistent under rules: you follow a style guide exactly, on task one and task four hundred Quick to learn specialized editing interfaces Methodical and patient — the small things are the job, not a distraction from it Reliable without supervision, on deadline Able to parse accents, dialects, mumbling, and overlapping speakers Nice to have Closed captioning, broadcast transcription, subtitling, or court reporting Linguistics, phonetics, or audio data annotation Waveform or caption tooling — Praat, ELAN, Subtitle Edit, or similar Speech-to-text QA or audio post-productionPrior annotation work against detailed guidelines Fast, accurate typing Why this one The audio you touch trains frontier speech systems built with leading research labs Fully remote, genuinely flexible — you choose when and how much Freelance autonomy, but the work has structure and a point to it Your captioning or linguistics background is the qualification here, not a detour from it Ongoing work and contract extensions as new projects open #J-18808-Ljbffr