About the role
Perit AI is hiring for Fixed-Term Audio Rating
About the Company
Perit AI builds AI and machine-learning training data for frontier labs. Its work spans audio, image, video, and reinforcement-learning environments, creating human-generated datasets that support the development of advanced models. This fully remote opportunity focuses on careful evaluation of AI voice-agent conversations.
Responsibilities:
- Listen closely to recorded conversations between people and AI voice agents, including fast, casual, or overlapping speech.
- Rate each agent turn for both the quality of its spoken response and the quality of its voice delivery.
- Write concise, specific justifications in your own words, identifying issues such as robotic-sounding sections, changes in voice identity, off-beat pacing, misplaced emphasis, or weak conversational responses.
- Apply sound judgment consistently and flag poor performance directly rather than overlooking issues.
Requirements
- No prior professional experience is required; training will be provided.
- Strong listening comprehension and a good ear for American English are essential.
- You should be able to explain ratings clearly and specifically in original writing.
- This role requires focused attention and is not suitable for multitasking or using AI to write rating justifications.
- The position is remote, freelance/contract-based, full-time, and lists a commitment of 30–50 hours per week as well as 8 hours per day.
- Monthly compensation is ₹20,000–₹30,000, depending on output.