Freelance Audio Contributor – US English

Job description

About the Role We're looking for native US English speaking university students to join our AI audio training pipeline as Freelance Audio Contributors. This is your chance to be on the front lines of how AI learns to understand, process, and generate human speech. You'll record conversational audio from provided scripts, feeding directly into the training loops of next-gen AI systems. Remote. Flexible. Project-based. Built for students. What You'll Do AI Voice Data Generation Record natural two-speaker conversations from provided scripts and prompts designed to train AI speech models Keep dialogue sounding natural and spontaneous to help AI learn authentic human conversation patterns Ensure recordings meet technical specs: minimum 16kHz sample rate, clear audio, minimal background noise Each file must be between 10 seconds and 1 minute in length What You Need Native US English speaker with a standard American accent Access to a quiet recording environment and a reliable headset/microphone Strong attention to detail and ability to follow written guidelines Ability to work independently and meet batch delivery deadlines Basic computer literacy (file management, audio uploads, spreadsheet tools) Prior experience in transcription, voice work, or data annotation is a plus but not required Why This Matters Every major AI breakthrough starts with high-quality training data. The models behind voice assistants, AI-generated speech, and conversational AI all depend on real human voices and real human judgment. This isn't just a side gig. You're directly contributing to how AI evolves. What's in it for You $20 per 60 minutes of recorded audio Fully remote, work from anywhere quiet Flexible hours that fit around your class schedule Hands-on experience inside a real AI data pipeline Understand how AI models are trained from the ground up Important Notes All audio data must be original and legally sourced Submitted data must not include any personally identifiable information (PII) of third parties without explicit consent Perle retains all intellectual property rights to collected data and associated deliverables Apply now and be part of the data layer that's shaping the future of AI! About Perle Perle is a data sourcing company dedicated to building high-quality, human-generated datasets that power the next Perle AI builds the foundational training data that powers the world's most advanced AI models. We partner with leading AI companies to source high-quality, human-generated datasets that make AI smarter, more natural, and more capable. Our contributor network spans the globe, and our data directly shapes how AI understands and generates human language.