Freelance Audio Conversation Rater

Perit AI; United States; Remote; USD 2,400 - 3,500; Part time

Benefits offered by this job

Remote work

Freelance

Flexible hours

Job summary

Perit AI builds AI and ML training data for frontier labs. You will listen to recorded conversations between a person and an AI voice agent, rate every turn on content and tone, and write a short justification in your own words.

This is a fully remote freelance role, requiring a good ear for American English and clear writing. Expect 30–50 hours per week, with compensation in INR monthly, and a flexible schedule.

Qualifications

• A good ear for American English.
• Real listening comprehension.
• You can follow fast, casual, overlapping speech without replaying it multiple times.
• Your own writing. AI-generated justifications are rejected.
• This is a writing job as much as a listening one.
• No prior experience required — we train you.

Responsibilities

• Listen to recorded conversations between a person and an AI voice agent.
• Rate every agent turn on what it said and on how it sounded.
• Write a short justification for each rating, in your own words.
• Call a weak response weak and note issues clearly.

Skills

American English listening

Strong writing

Attention to detail

Job description

Perit builds AI and ML training data for frontier labs. We work across audio, image, video, and reinforcement learning environments, producing the high-quality, human-generated datasets that the next generation of models is trained on. You listen to recorded conversations between a person and an AI voice agent, and rate every agent turn on two things — what it said, and how it sounded — each with a short written justification in your own words. Not for you if you were planning to run the task through ChatGPT, or you want something to half-watch while doing other work. The bar is high, and it's checked.

What you'll be doing

• Listening to recorded conversations between a person and an AI voice agent.
• Rating every agent turn on what it said and on how it sounded.
• Writing a short justification for each rating, in your own words.
• Calling a weak response weak — the most common failure is noticing a problem and scoring it generously anyway.

What you'll need

A good ear for American English. You notice when a voice slips: a robotic patch mid-sentence, a shift in voice identity between turns, pacing that's a beat off, stress landing on the wrong word. Real listening comprehension. You can follow fast, casual, overlapping speech without replaying it three times. Opinions you can defend. Most of this job is writing one or two sentences on why a response was weak — specifically. "It was okay" is worth nothing to us. Your own writing. AI-generated justifications are obvious and we reject them. This is a writing job as much as a listening one. No prior experience required — we train you. 30–50 hours a week, fully remote.

Pay

₹20,000 to ₹30,000 a month, depending on output. Freelance, 30–50 hours a week, fully remote.

Listed rate: $210–315/month.

Perit is a Y Combinator-backed speech-data company. Work is remote;