Food Video Highlights Annotation Contractor

About OpenTrain

OpenTrain AI is the hiring and contracting organization for this opportunity and the #1 platform for finding and building careers in AI training and data labeling. Creating an OpenTrain account is free, and eligible contributors can apply in minutes.

This contractor role offers a chance to contribute to cutting-edge AI systems while working on food and cooking video data from anywhere in the world.

• Contractor opportunity
• Worldwide participation
• English-language project
• Pay: $0.05 per completed label

About AI Training and Video Annotation

AI training is the human side of building artificial intelligence. People prepare and review examples that help modern models understand images, video, language, and actions.

In video annotation, contributors identify meaningful moments and describe what is visibly happening. Your careful labels can help AI systems recognize cooking activities and other events in visual content.

• Work with real video examples used for AI training
• Build experience in data labeling and visual understanding
• Contribute to how AI models interpret actions and objects

The Role

As a Food Video Highlights Annotation Contractor, you will support a video moment retrieval project in the food domain. You will match natural-language queries to relevant moments in videos, mark accurate temporal boundaries, classify the action and objects present, and create visual proxy descriptions for CLIP training.

This is an entry-level role for candidates with a graduate degree and an understanding of food videos. The work requires close attention to timing, visual detail, and the difference between what a video visibly shows and what you may infer from it.

• Data type: Video
• Label types: Action recognition and classification
• Platform: Label Studio
• Language: English
• Experience level: Entry level

What You’ll Do

You will review food videos against natural-language queries and identify every segment that matches. Each annotation should include precise timing, relevant classification metadata, and a clear description grounded in visible content.

• Read and interpret each natural-language query before reviewing the video
• Watch videos to find all matching action or event segments
• Mark start and end times within 0.5 seconds when boundaries are clear
• Classify the action type and objects present
• Write a visual proxy description for CLIP training
• Describe what you see rather than relying on assumptions
• Use high, medium, or low confidence based on boundary certainty

Food Video Actions You May Label

The project covers a broad range of food preparation, presentation, and reaction moments. Labels should reflect the specific visible action and its relevant objects.

• Chopping ingredients, mise en place, and mixing ingredients
• Kneading dough, sautéing, stirring the pot, and deglazing
• Tasting food, adding seasoning, boiling, grilling, and baking
• Plating the dish, garnishing, and sauce drizzle
• Recipe introduction, finished dish reveal, and eating reaction

Quality Standards

High-quality annotations are precise, complete, and visually grounded. Avoid selecting segments that are too short to capture the action or so long that they include unrelated content.

• Match the query specifically, not broadly or ambiguously
• Mark all matching segments instead of stopping at the first result
• Aim for boundaries within 0.5 seconds of the actual moment
• Use a 1-to-2-second range for medium-confidence boundaries when needed
• Flag low confidence when the action or boundaries are unclear
• Write specific descriptions such as a person in a white chef coat slicing red tomatoes
• Avoid vague descriptions such as cooking or delicious food

Requirements

This is an entry-level project, but candidates must meet the stated education and subject understanding requirements. No additional skills were specified for the role.

• Graduate degree
• Understanding of food videos
• English-language proficiency sufficient to interpret queries and write descriptions
• Careful attention to visual detail and timing
• Ability to complete labeling work in Label Studio

How to Apply

Create a free OpenTrain account and apply through OpenTrain AI for consideration. If selected, you will work as a contractor labeling food video moments and helping improve the data used to train AI systems.

• Apply online through OpenTrain
• Review project instructions before beginning
• Complete labels according to the timing, classification, and description guidelines
• Earn $0.05 per completed label