About OpenTrain AI

OpenTrain AI is the #1 platform for finding and building careers in AI training and data labeling. We connect skilled contributors with real, paid work that helps shape how state-of-the-art AI systems learn and behave.

About AI Training Work

AI training (also called data labeling or annotation) is the human work behind modern machine learning: people prepare and check examples—audio, text, images—that models learn from. These roles are often remote, flexible, and accessible to people with language skills, attention to detail, or domain experience.

This project supports next-generation speech recognition and language-learning models by producing precise phoneme- and word-level transcripts of spoken French for acoustic and ASR training.

The Role

You will convert long French audio clips (about 1–2 minutes each) into word- and phoneme-level transcripts using ARPABET notation in a web-based annotation tool. Tasks are delivered in rolling batches; your work will pass through a peer QA loop before final acceptance.

• Role type: Contractor, part-time, ongoing production work.
• Location: Global — work remotely from any country.
• Language: Native-level French required; conversational English for instructions and QA.

What You’ll Do

Produce accurate word-level alignments and phoneme-level transcriptions using ARPABET for French speech. Detect and label reading errors and disfluencies, and revise annotations based on peer QA feedback.

• Transcribe long (1–2 minute) French audio clips at word and phoneme levels.
• Apply ARPABET phonemic notation consistently and accurately.
• Identify and label reading errors: substitutions, omissions, insertions, misreads.
• Identify and label disfluencies: filled pauses, repetitions, false starts, self-corrections.
• Work in a web-based annotation tool and participate in peer QA loops.

Requirements

You must meet all required qualifications and be ready to start within one week of acceptance. A qualification quiz must be passed before you receive production batches.

• Native-level French with ability to judge fine-grained pronunciation and reading accuracy.
• Intermediate to expert proficiency with ARPABET phonemic transcription for French.
• Demonstrated experience in phonetics/linguistics, ASR dataset transcription, phoneme-level labeling, or child-speech transcription.
• Proven ability to produce consistent word-level alignment and phoneme-level transcription.
• Skilled at detecting reading errors and labeling speech disfluencies with high accuracy.
• Availability to commit 20+ hours per week on an ongoing basis.
• Able to start within 1 week of acceptance and pass a qualification quiz.
• Comfortable working in a web-based annotation tool and following peer-QA feedback.

Compensation & Schedule

Pay is listed as $18–$20 USD per hour. Work is paid for approved tasks; higher throughput and accuracy typically increase overall earnings. You will receive production batches on a rolling basis and are paid per approved work according to the project's payment system.

• Hourly rate: $18–$20 USD (PAY_PER_HOUR).
• Time requirement: 20+ hours per week; schedule is flexible within that commitment.
• Payment is contingent on approved tasks and peer QA acceptance.

How the Process Works

Apply through OpenTrain AI, complete the qualification quiz, and begin working on trial batches. After qualification, you'll receive rolling annotation batches and submit work via the web-based tool. Your submissions will be reviewed in a peer QA loop before final acceptance and payment.

• Start: must be available to begin within one week of acceptance.
• Qualification: pass a required quiz to access production work.
• Delivery: rolling batches; deliver accurate, consistent transcriptions to receive approval and payment.