AI Trainer – Audio (English) – Speech Transcription & Timestamp Alignment

About the role

We are seeking detail‑oriented English speakers to improve the quality of conversational speech data used to train AI models. You will listen to natural US English audio, edit automatically generated transcripts, and align timestamps so the text matches the speech precisely.

Key responsibilities

• Listen to naturalistic conversational audio files and review ASR‑generated transcripts.
• Correct transcription errors according to client‑provided conventions and formatting rules.
• Adjust start and end timestamps for each segment to ensure accurate audio‑text alignment.
• Maintain high accuracy and consistency across all assigned files.
• Follow detailed instructions and meet required deadlines.

Required profile

• Near‑native listening proficiency in English (Auditory C1 level or higher).
• Written English proficiency at B2 level or above.
• Excellent attention to detail and ability to work quickly and precisely.
• Comfort with casual speech, disfluencies, slang, and regional accents.
• Ability to work independently and consistently meet deadlines.

Required skills

• Experience in transcription, subtitling, annotation, or speech‑to‑text editing.
• Familiarity with audio‑text alignment workflows.
• Strong grammar, punctuation, and proofreading abilities.