Arabic Audio Transcription QA Specialist

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help contributors discover projects, consolidate their work history, and build a durable freelance career teaching AI — all from a single, free account.

OpenTrain is the hiring and contracting organization for this role. If selected you will work directly with OpenTrain on a contract, part-time basis.

About AI training and why it matters

AI training (data labeling/annotation) is the human side of building modern AI: people create, correct, and evaluate examples that models learn from. These roles are often fully remote, flexible, and accessible to contributors with language skills, attention to detail, or domain expertise.

By working on evaluation-grade speech data you will directly shape how Arabic-speaking voice and transcription systems perform in the real world.

The role

OpenTrain is hiring an Arabic Audio Transcription QA Specialist to review multi-channel conversations for an evaluation-grade speech dataset. You will validate audio fidelity, confirm channel isolation and quiet recording conditions, check transcription accuracy against strict verbatim rules, verify diarization and speaker timestamps, and audit metadata, privacy, and content quality across human-verified annotations.

This is a part-time contractor role (20+ hours/week) and is open to applicants worldwide who meet the language and skill requirements. Work focuses on audio data, transcription review, and evaluation rating of annotations.

• Data type: multi-channel AUDIO
• Labeling tasks: TRANSCRIPTION and EVALUATION_RATING
• Employment: CONTRACTOR, PART_TIME
• Time requirement: 20+ hours per week

What you'll do

• Listen to multi-channel audio and flag fidelity issues such as bleed, clipping, low gain, and excessive background noise.
• Validate human-corrected transcriptions for verbatim accuracy, including overlaps, interruptions, false starts, and disfluencies.
• Verify turn-level and word-level timestamps, speaker diarization, and speaker-identification labels for consistency.
• Inspect JSON-formatted metadata and annotation logic for correctness and completeness.
• Audit sessions for demographic markers, conversational tags, personally identifiable information (PII), toxic or unsafe content, and other privacy concerns.
• Assign clear pass/fail or agree/disagree outcomes and leave actionable review feedback when annotations need correction.

Requirements

• Native-level fluency in Arabic (required) for verbatim transcription review.
• Proven ability to judge audio fidelity issues (bleed, clipping, low gain, background noise).
• Experience reviewing multi-speaker conversations and diarization quality.
• Careful attention to verbatim transcription rules and timestamp precision (turn-level and word-level).
• Comfort reading structured annotation data (JSON) and making quality-control decisions.
• Ability to audit PII, toxic content, and metadata for safety and privacy.
• Availability for 20+ hours per week; work is remote and contract-based.

Helpful background

• Formal or practical experience in linguistics, phonetics, or speech analysis.
• Professional transcription experience with strict accuracy targets for natural, unnormalized speech.
• Previous work in speech QA, annotation, or AI training workflows.