Freelance Opportunity: Arabic Language Transcription Specialist

Project Description

Uber AI Solutions is seeking detail-oriented specialists for transcription to support a large-scale generative AI training program. In this engagement, you will transcribe and annotate audio files with accuracy, capturing every linguistic nuance exactly as spoken.

Language: Arabic

We are looking for native speakers for Arabic transcription. Applicants must be able to accurately transcribe any of the following varieties:

• Modern Standard Arabic (ar-001 | ar-MSA)
• Egyptian (ar-EG)
• Maghrebi (ar-MA | ar-DZ | ar-TN | ar-LY)
• Levantine (ar-LB | ar-SY | ar-JO | ar-PS)
• Gulf / Khaleeji (ar-SA | ar-AE | ar-KW | ar-QA | ar-OM | ar-BH)

What you’ll work on

• Transcription: Transcribe audio with 98% accuracy, capturing every disfluency, filler word, false start, and stutter exactly as heard.
• Precision Timestamping: Align text segments to the audio waveform with millisecond precision (max gap <500ms).
• Speaker Identification: Accurately identify and label speakers in multi-speaker audio files (2–8 interlocutors).
• Tagging and Annotation: Apply correct tags for non-speech events like (laughs) or (applause) and unintelligible segments.

Skills and Qualifications

• Native-level fluency: You must be a native speaker of the specific regional dialect.
• Attention to Detail: You can distinguish between "clean" speech and "verbatim" speech (e.g., capturing the exact phonetic realization of words based on regional pronunciation).
• Tech Savvy: Comfortable navigating web-based annotation tools.

Engagement Details

• Volume: Steady task flow available for high-quality contributors.
• Flexibility: Work on your own schedule, provided quality and deadline standards are met.