Job Description

Voice Annotator: If you know any of the languages listed below and have worked on AI training projects before, we'd love to hear from you.

Location: Remote (India) / Bengaluru

Engagement: Project-based, freelance

Languages required (any one/multiple): Hindi, Bengali, Marathi, Gujarati, Tamil, Telugu, Kannada, Malayalam, Punjabi, Indian English, Urdu, Odia, Assamese, Konkani, Sindhi, Kashmiri, Nepali, Manipuri (Meitei), Bodo, Dogri, Santali, Sanskrit, Maithili

About the role:

Arctic Engines builds voice and speech data for frontier AI labs, training conversational AI to sound natural across Indian languages. We're hiring native or highly fluent speakers of any of the languages listed above to listen to and evaluate voice samples, judging whether they sound like a real, natural speaker of that language or fall short. This is focused listening work. You will not be writing scripts or performing voiceovers. Your job is to catch what a native ear catches: an accent that's slightly off, a pause in the wrong place, a sentence that sounds stiff instead of conversational.

What you'll do

- Listen to short AI-generated voice samples in your language and judge how natural and native-sounding they are.
- Flag issues with accent, pronunciation, pacing, tone, or audio clarity.




- Score samples consistently against a shared rubric, so your judgments line up with other evaluators.
- Provide short, specific written notes when rejecting or flagging a sample.
- Take part in periodic calibration sessions to keep scoring aligned across the team.

What we're looking for

- Native or highly fluent speaker of one or more of the languages listed above, comfortable with regional accent variation.
- A positive ear: you notice when speech sounds robotic, mispronounced, or just "off", even if you can't always explain why in technical terms.
- Comfortable with repetitive, detail-focused listening work over sustained sessions.
- Reliable internet connection and a quiet space to work with headphones.
- Prior experience in audio QC, voice-over, linguistics, or language teaching is a plus, not required.

Hiring process

- Application review
- Short scoring exercise: you'll evaluate a small batch of sample voices against our evaluation rubric, so we can see how your judgment aligns with the standard before onboarding.
- Offer and onboarding, including a calibration session with the QC lead

Compensation:

- Hourly, annotators who complete tasks consistently and on time can up to INR 10,000 to 20,000.

📌 Voice Annotator (India)
🏢 Arctic Engines
📍 India