Video Annotator

Role Overview

We seek meticulous Video Annotators for a Vision-Language-Action dataset initiative. This role entails viewing short video clips depicting physical activities and segmenting each clip into meaningful events. Annotators must then generate clear and precise natural-language instructions that accurately describe the actions within each segment.

Key Responsibilities

• Watch each assigned video completely prior to annotation to understand the task context.
• Divide videos into distinct, logical steps or events based on visible action boundaries.
• Create clear, concise, and grammatically accurate step-by-step instructions for each video segment.
• Ensure the annotations are action-focused, unambiguous, and consistent in tone, tense, and phrasing throughout the dataset.
• Mark correct start and end timestamps for each annotated segment.
• Adhere to annotation guidelines, taxonomy structures, and style instructions detailed by the project team.
• Report any videos that are unclear, corrupted, mislabelled, or inappropriate for annotation.
• Engage in calibration exercises and apply reviewer feedback to enhance quality and consistency of work.
• Consistently meet productivity targets for both quantity and accuracy within daily and weekly deadlines.
• Keep all project information and materials confidential.

Mandatory Qualifications

• Fluent English at C1 CEFR level or above, with strong grammar and vocabulary skills essential for writing clarity.
• Exceptional attention to detail and ability to detect subtle visual changes.
• Proficiency in written communication to describe activities clearly and succinctly.
• Comfortable using video annotation software and tools; training will be provided.
• Ability to work independently while following detailed instructions and maintaining consistency in repetitive tasks.
• Access to a reliable internet connection and a computer with a modern web browser.
• Basic computer skills including file management, web-based applications, and spreadsheet use.

Preferred Skills and Experience

• Experience in data annotation, labeling, transcription, or content moderation.
• Understanding of AI and machine learning concepts, especially in computer vision, NLP, or robotics datasets.
• Experience creating instructional or procedural text such as guides, SOPs, or recipes.
• Familiarity with annotation platforms like CVAT, Label Studio, Scale, or similar.

Attributes We Value

• Strong observational skills and passion for writing precision.
• Capability to watch activities and transform them into detailed written instructions suitable for human or machine use.
• Self-motivated with an aptitude for focused, repetitive work and embracing iterative feedback.

Selection Process

• An English proficiency test confirming C1 level.
• Assessment on annotation skills including video segmentation and description writing.
• Interview and calibration session.
• Onboarding along with comprehensive guideline training.