Associate Analyst IRC307505

About the Role

We are seeking a detail-oriented, energetic, and physically active candidate to join our team in a dynamic dual-function role: Egocentric Data Collection and Video Content Annotation & Quality Review. In this position, you will play a hands-on role in capturing and curating high-value training datasets for cutting-edge Physical AI and robotics models.

Key Responsibilities

Egocentric Data Collection

• Collect high-quality, first-person (egocentric) video footage across diverse real-world environments (e.g., retail stores, warehouses, hospitality venues, workshops, and lab settings) using advanced head-mounted hardware, including Meta Quest 3 VR headsets.
• Perform realistic, structured, and multi-step physical activities while wearing the Meta Quest 3 headset to capture clear first-person video data.
• Execute a wide range of natural hand-object interactions, including picking, placing, reaching, bending, turning, carrying, packing, sorting, organizing, scanning, loading, and trolley transport.
• Operate across diverse, real-world scenes such as retail shelving areas, hotel guest rooms, warehouse packing benches, assembly stations, and laboratory setups.
• Ensure all tasks demonstrate dynamic whole-body movement and clear hand visibility without blocking camera views.

Video Annotation & Review

• Review and analyze first-person recordings capturing human and robotic task execution.
• Video Segmentation & Timestamping: Carefully review raw video footage to identify key actions, events, and task transitions.
• Split/cut long video files into micro-segments with exact start and end timestamps (down to the millisecond/frame level).
• Classify video segments according to specific predefined categories, action taxonomies, or labeling guidelines.
• Action-Level Captioning & Text Annotation: Write clear, concise, and objective natural language sentences describing the exact task or action taking place within each segment (e.g., “Person picks up the blue wrench with their right hand and turns the bolt counter-clockwise”).
• Follow strict guidelines on grammar, tense, vocabulary, and formatting to maintain consistency across datasets.
• Ensure descriptions capture context, objects used, sub-tasks, and spatial/temporal relationships accurately.

Data Integrity & Intent Verification

• Identify and flag fraudulent intent, artificial motion, or non-authentic behavior in recorded data to ensure videos capture genuine, purposeful task execution.
• Verify that all recorded content strictly adheres to established quality, technical, and privacy standards.
• Maintain high accuracy and consistency across large volumes of video data.
• Perform self-audits and peer reviews to detect and correct misaligned timestamps or ambiguous task descriptions.
• Adhere strictly to project-specific annotation guidelines and edge-case instructions provided by AI/ML engineers or Team Leads.

Daily Output & Collaboration

• Meet daily and weekly quota targets for video hours processed and segments annotated.
• Collaborate with Quality Analysts (QA) and Data Operations Managers to resolve edge cases and refine descriptive guidelines.

Requirements & Tech Stack

• Meta Quest 3 VR headsets
• Familiarity with video labeling/editing software
• Understanding of frame rates (FPS) and timestamps
• Basic computer proficiency, familiarity with MS Office / Google Workspace
• Ability to learn new review and annotation tools quickly

Qualifications

• Strong interest in Physical AI and Robotics.
• Solid computer literacy.
• Sharp analytical judgment.
• Exceptional eye for detail.
• Strong written and verbal English skills with precise grammar and concise descriptive capabilities.
• Ability to spot micro-movements, subtle activity shifts, and precise cut points in fast-moving videos.
• Ability to break down complex human activities into step-by-step sequential tasks.
• Ability to sustain focus and accuracy across high volumes of repetitive review work.
• Sound judgment and critical thinking to evaluate intent, context, and authenticity in video content.
• Comfortable decoding and applying detailed written instructions and rubrics.
• Able to describe video content and review findings with appropriate grammar, vocabulary, and structure.
• Typing Speed: 30-35 words per minute with 90% accuracy.
• Ability to work effectively under supervision, manage time, and meet tight deadlines in a fast-paced environment.
• Comfort working with large-scale datasets and review queues.
• High integrity and discretion when handling sensitive or private footage.