(Speech and Fluency Analysis)

We are looking to find a skilled Speech Disfluency and Repair Event Annotation Specialist to join our growing team.

You will annotate disfluency features in spontaneous speech, including hesitations, fillers, repetitions and repair sequences. These annotations support models that learn natural spoken language behaviour rather than scripted delivery.

Responsibilities:

• Listen to spontaneous speech and label filled pauses, silent pauses above threshold, repetitions, restarts and self-corrections

• Distinguish natural rhythmic pausing from true disfluency according to guidelines

• Mark the span of repair segments where the speaker revises or corrects an utterance

• Provide consistent tagging of filler tokens such as “uh” and “um” across speakers and languages

• Flag cases where overlap, poor audio quality or noise obscure accurate tagging

Requirements:

• Patience and precision in close listening tasks

• Familiarity with concepts of fluency, disfluency and spontaneous speech

• Ability to work with annotation tools that support token-level or word-level alignment

Background:

•  Linguistics, conversation analysis, clinical speech work or call review

•  Experience working with unscripted or conversational audio

Sample Test Task:

• Given a one-minute unscripted speech clip, annotate all disfluency events, classify each by type and provide a short explanation for any ambiguous cases.

Application Requirements:

• Short note explaining how you distinguish natural pauses from disfluent hesitations

• Optional examples of past work with conversational or spontaneous speech data

• Completion of the test task

Compensation

• Per minute of conversational audio or per batch of transcribed clips, depending on project.