Egocentric Video Annotation Specialist

About OpenTrain

OpenTrain AI is the organization hiring and contracting for this role. OpenTrain is the #1 platform for finding and building careers in AI training and data labeling, helping contributors discover projects, build profiles, and apply to cutting-edge work. Creating an OpenTrain account is free.

About AI Training Work

AI systems learn from examples prepared and reviewed by people. In video annotation, contributors identify actions, describe what happens, and align labels precisely with the underlying footage so models can better understand human activity and motion.

• Remote work completed with a computer, internet connection, and the OpenTrain platform
• Flexible, part-time contract work in a fast-growing technology field
• Your annotations help evaluate and improve how AI interprets real-world video

The Role

As an Egocentric Video Annotation Specialist, you will label first-person videos showing human actions and motion trajectories. The project uses a three-tier captioning scheme and requires careful adherence to detailed guidelines, strong written English, and high-quality temporal annotation.

The work is intermediate-level and is available to contractors in India and the Philippines. The rate is $8 USD per hour, with an expected commitment of at least 20 hours per week. A pilot is planned for February 3, with potential to scale from 1,000 to as many as 10,000 video hours.

• Pay: $8 USD per hour
• Engagement: Part-time contractor
• Location: India or the Philippines
• Time requirement: 20+ hours per week
• Data type: Video
• Labels: Text generation, action recognition, and tracking

What You’ll Do

You will work with pre-labeled first-person video and improve or create detailed annotations. Each tier must remain grounded in what is visibly shown, without guessing about a person’s intent.

Your work will be assessed against description accuracy, description completeness, and timestamp precision, with a target quality level of 95% or higher.

• Write one high-level video summary of one to two sentences without timestamps
• Improve pre-annotated action-level segments with clear start and end timestamps
• Write atomic action labels in verb plus object format, such as “grasp cup” or “place lid”
• Create trajectory-level descriptions from scratch using sub-second segments
• Describe visible motion at the body-part level, including the left hand, right hand, or torso
• Use overlapping segments when needed to represent simultaneous limb movements
• Maintain precise temporal alignment and follow all project guidelines

Requirements

This role requires C1 English proficiency or higher and the ability to write precise, natural descriptions. You should have prior experience with video annotation, especially action segmentation or temporal labeling, and be comfortable working with detailed customer-provided instructions.

Applicants should be able to complete a short qualification check before starting and support the February pilot. Familiarity with OpenTrain or an equivalent annotation tool is expected.

• C1 English proficiency or higher
• Prior experience with video annotation, action segmentation, or temporal labels
• Experience timestamping actions with tight start and end alignment
• Ability to write observable verb plus object action labels
• Ability to create body-part-level trajectory descriptions using segments shorter than one second
• Familiarity with OpenTrain or equivalent annotation tools
• A demonstrated record of high-quality work, targeting 95% or higher accuracy and low rework
• Availability for the pilot beginning February 3 and potential subsequent scale-up
• Availability for at least 20 hours per week

Who Should Apply

This opportunity suits experienced video annotators and quality-focused review specialists who notice fine-grained movement, distinguish observable actions from assumptions, and can consistently apply a structured captioning system. It may be a strong fit if you enjoy detailed visual work and want to contribute directly to the development of AI video understanding.

• Video annotators with action-level labeling experience
• Reviewers accustomed to precision, completeness, and low rework
• Contributors comfortable describing hand, torso, and other body-part movements
• Writers with strong English comprehension and careful attention to detail

How to Apply Through OpenTrain

Create or use your free OpenTrain account to apply. During screening, be prepared to confirm your C1 English level, describe relevant video annotation experience, and complete the required qualification check before beginning project work.

• Review the project requirements carefully
• Confirm your English proficiency and relevant experience
• Indicate that you can support at least 20 hours per week
• Be ready to work in the OpenTrain annotation tool and follow detailed guidelines