About the Role
We are seeking detail-oriented Non-Music Generalist Annotators to support an audio captioning project involving English speech transcription, speaker identification, voice description, sound-event annotation, and audio quality review. Candidates should have excellent English writing skills, strong listening abilities, and the ability to describe voices and sounds accurately based on audio evidence. Backgrounds in linguistics, voice analysis or coaching, and audio production are preferred.

Key Responsibilities
• Writing: Write natural, fluent English descriptions using precise and concise vocabulary to characterize voices and sounds.
• Speaker Tracking: Attribute speech to the correct speaker, including during overlaps, interruptions, and group conversations. Identify when a new speaker enters.
• Transcription: Accurately transcribe English speech, including repetitions, fillers, and names. Recognize different English accents and identify speech in other languages.
• Voice Description: Describe voice characteristics and changes in pitch, voice quality, loudness, pace, emotion, and intonation. Compare speakers based on audible characteristics rather than the meaning of their words.
• Non-Speech Vocalizations: Identify, distinguish, count, and timestamp vocal sounds such as laughter, coughing, sneezing, sobbing, gasping, and breathing.
• Environmental Sound Annotation: Identify and describe non-speech sounds by texture, onset, distance, and relative loudness.



Distinguish individual sound events from continuous background ambience.
• Sound Prominence: Identify prominent and masked sounds and track changes in their relative prominence, including when music begins, ends, or is replaced.
• Scene Authenticity: Assess whether voices and sounds belong to the original recorded scene or appear to be added, such as narration, playback, or background music.
• Timing Accuracy: Precisely identify event onsets and offsets, speaker changes, overlapping speech, and gradual transitions.
• Machine-Generated Draft Review: Review machine-generated annotations for missing, invented, mistimed, or incorrectly attributed content. Base corrections on audible evidence rather than video content, subject matter, or the draft itself. Qualifications and Skills
• Bachelor’s degree or above.
• Background in linguistics, voice analysis or coaching, or audio production is preferred.
• Excellent English comprehension and natural, native-level English writing ability.
• Strong auditory attention and the ability to distinguish speakers, vocalizations, and environmental sounds.
• Ability to transcribe speech faithfully and produce clear, accurate audio descriptions.
• Robust attention to detail, especially for timing, speaker attribution, overlapping speech, and sound-event identification.

How to Apply
Submit your application with an updated CV highlighting your relevant qualifications, language skills, and experience in linguistics, voice analysis, transcription, or audio production.

📌 Non-Music Generalist Annotator (India)
🏢 Innodata
📍 India