Andromeda Sadiradra - Audio Rater - Generalist - English (Australia)

Welo is an AI services company focused on training and evaluating artificial intelligence systems. The Welo Data team specializes in gathering human feedback to improve AI performance across various applications. This role sits at the heart of that work—you'll be assessing how well AI voice agents communicate in real conversations.

The Position

As an Audio Rater, you'll evaluate recorded exchanges between users and AI voice agents. The work involves listening closely to these interactions and scoring the agent's responses on two separate dimensions using a 1–5 scale. Content quality measures whether responses are helpful, relevant, and appropriately tailored. Prosody evaluates the technical delivery—does the voice sound natural, well-paced, and emotionally expressive? You'll provide brief written explanations alongside each score, capturing the reasoning behind your ratings.

About You

You have sharp listening skills and genuine attention to detail. You're someone who can make consistent judgments and articulate them clearly in writing. You approach evaluation tasks methodically, applying the same standards across multiple samples. Reliability matters; you're comfortable committing to a regular schedule of around 10 hours per week. You're fluent in Australian English and attuned to the nuances of how language is spoken in that context.

What We Offer

The role begins as soon as possible and runs as a freelance engagement at 27.30 per hour. You'll work entirely remotely from Australia, with the flexibility to fit approximately 10 hours per week into your schedule. This is an excellent opportunity if you're interested in AI development and want to contribute directly to making voice agents sound more natural and responsive.