Bilingual Audio Contributor (Indian Languages/English)

You will create and review speech data used to train and evaluate AI systems. The work includes recording scripted prompts and free conversation in an Indian language and English, including natural code-switched speech that reflects real bilingual communication.

You will also review transcripts, correct errors, and flag recordings where the audio does not match the written text. Recordings must follow the project specifications and be clear enough for use in AI training.

• Record scripted and conversational speech in an Indian language and English.
• Create natural recordings that switch between an Indian language and English.
• Review and correct speech transcripts.
• Identify audio that does not match its transcript.
• Follow recording instructions and maintain clean, usable audio.

What it pays and takes

This is a remote, asynchronous, part-time independent contractor role with ongoing project work. You can adjust your schedule from week to week, subject to the required weekly availability.

• Pay: $20 to $40 per hour, paid weekly through Stripe.
• Availability: 20 or more hours per week, with no stated maximum.
• Location: Remote and open to candidates in the listed countries.
• Language: Professional fluency in one or more Indian languages, such as Hindi, Tamil, Telugu, Bengali, Marathi, or Kannada, plus full professional fluency in English.
• Equipment: A quiet recording environment and a microphone that produces clean audio.
• Requirements: Comfort recording your own speech, following prompt specifications precisely, and spotting small transcription errors.
• Helpful experience: Voice-over, podcasting, audio production, transcription, data annotation, or fluency in more than one Indian language.

How it works

Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training work is the human side of building artificial intelligence. People record, label, review, and evaluate examples so AI systems can better understand speech and produce useful results.