Mizo Language Quality Analyst Specialist

Job Title: Quality Analyst (QA) – Specialist for Mizo language Job Description: We are seeking detail-oriented Quality Analyst (QA) to evaluate audio data. The role involves reviewing audio samples and ensuring they meet linguistic, technical, and quality standards. Key Responsibilities: - Ensure the audio is in Mizo language. - Confirm that the speaker is a native speaker of Mizo language.| - Identify non-native accents or unnatural speech patterns. - Verify that spoken sentences are grammatically correct, natural, and complete. - Ensure the audio is clear and free from major distortions such as noise, clipping, or inaudibility. QA TASK INSTRUCTIONS The QA must choose ONE of the following if applicable: 1) File Issues (Mandatory Check) - If the file has 0 KB size (empty file), choose this option. - If the file has > 0 KB size, but no human speech or vocal sounds (e.g., breathing, coughing, laughing), choose this option. 2) Language Verification - If the audio is in Mizo language, mark as correct. - If the audio is not in Mizo language, select this and specify: - actual_language_detected 3) Fluency & Nativeness - Speech is natural, fluent, and spoken by the native speaker. - Speaker has a non-native accent or unnatural pronunciation. - Speech contains frequent pauses, hesitations, repetitions, or unnatural flow. 4) Grammar & Sentence Completeness - Sentence is complete, natural, and grammatically correct. - Sentence contains grammar mistakes. - Sentence is cut off or unfinished. 5) Audio Quality - Clear and audible with no major issues. - Presence of noticeable environmental noise. - Audio is crackling, clipped, or broken. - Speech is too quiet to understand clearly. - Audio is too loud or distorted due to clipping. - More than one voice speaking at the same time. 6) Speech Style Issues - machine_generated_speech: Audio sounds like text-to-speech (TTS). - reading_style_speech: Speech sounds overly scripted, monotone, or unnatural (not conversational). 7) Speaker Validation - single_speaker: Only one speaker is present. - multiple_speakers: More than one speaker detected in the same audio. 9) Dialect Verification (If applicable) - Matches the required dialect. - Does not match; specify: detected_dialect Requirements: - Native or near-native fluency in Mizo language - Strong listening and linguistic judgment skills - Attention to detail and ability to follow strict guidelines - Prior experience in audio QA is preferred

Project ID: 40394768