(Action Recognition / Motion Annotation / Temporal Segmentation)

We are looking to find a skilled Video Action Labelling Analyst to join our growing team.

You will annotate physical actions in video content including sports, stunts, fights, and high-motion sequences. The role focuses on identifying action intervals, classifying action types, and capturing intensity and temporal structure for action recognition models.

Responsibilities:

• Review video clips and identify action intervals with accurate start and end timestamps

• Label each action using a predefined taxonomy such as run, jump, punch, fall, dodge, tackle, slide or climb

• Annotate motion intensity according to a defined scale and record the number of active participants

• Identify key frames such as peak motion, impact moments or key transitions

• Flag cases where visibility, occlusion or blur create ambiguity, and document note

• Deliver structured annotation outputs aligned with Beatpulse motion schemas

Requirements:

•  Strong visual tracking skills for fast, complex movement

• Ability to use frame-accurate annotation and timeline tools

• Consistent application of predefined taxonomies and rules

Background:

• Sports analysis, choreography review, martial arts evaluation, stunt breakdowns or similar motion-focused work

• Editing or reviewing action footage for film, gaming, social content or coaching

• Ability to sustain long periods of precise visual work

Sample Test Task:

• Given a 45 second chase sequence with brief physical interaction, identify all action segments, label action type, timestamps, intensity level, and number of participants, and note any ambiguous areas.

Application Requirements:

• Short note describing how you define separate actions in a continuous sequence

• Examples of any past motion analysis or breakdown work (optional)

• Completion of the test task

Compensation

• Per action segment or per annotated minute, with adjustments for dense or complex scenes.