Posted: Feb 2, 2026
Egocentric Video Annotation - C1 English Required
Total Budget: $9,600OtherVideoPay Per Hour
Overview
Dataset
Labeling Details
Hiring
Budget
Client
Labeling Overview
Hi Team, Hope you are doing well. We are reaching out to explore potential collaboration on an upcoming pilot project and would like to get your inputs on feasibility availability and indicative pricing. Please note this outreach is for estimation and planning purposes only. No commitment is expected at this stage. Project Overview Project Name: Human Generated Egocentric Annotations [Action and Trajectory Level Annotations] Target Completion Date: End of February Objective: Benchmark the model capability to generate multiple levels of captions Task Description: Dense video annotations Volume: 1000 hours of video with potential to scale up to 10000 hours Accuracy Rate: 95%+ Acceptance Criteria: Annotation must meet following requirements, Description accuracy Description completeness Timestamp precision Annotation Tool: SuperAnnotate platform Quote Needed By. January 30 2026 Task Summary This project involves annotating first person egocentric videos following detailed customer provided guidelines. The work requires high attention to detail and strong English comprehension. Full task instructions and expectations are covered in the attached guideline. Pilot and Scale We are planning an initial pilot phase in February. Based on pilot performance the project may scale further. Information Requested: After reviewing the attached guidelines please share the following. Indicative hourly rates in USD for the required roles. Required skill level and experience to deliver high quality output. Recommended number of resources for the pilot phase. Approximate number of resources needed if the project scales to full production. Any early assumptions on annotation time per hour of video if available. Expected ramp up time and any operational constraints. Reference Material Annotation guidelines are attached for your review. Given the tight RFQ timeline we would appreciate your response as soon as possible. Thank you for your time and support. Looking forward to your inputs. Best Regards, ==== Thought for 42s Job Name: Egocentric Video Annotation (Tier 1–3 Captions) — C1 English, SuperAnnotate Dataset Description (5–8 words): Egocentric videos with multi-tier captions Data Type (select one): Video Subject Matter/Industry (5–8 words): Human actions and motion trajectories in video Pre-labeled Data (Yes/No): Yes Labeling Software: Other Label Types (multi-select): Action Recognition, Tracking, Text Generation Labeling Overview Qualifications / requirements: We’re looking for experienced video annotators (and/or QA/review specialists) with C1-level English who can follow detailed guidelines and deliver high-precision timestamps and high-accuracy descriptions (target 95%+). You should be comfortable working in the SuperAnnotate tool and have prior experience with action-level video labeling and quality-focused review workflows. What you’ll be doing: You will annotate first-person (egocentric) videos using a 3-tier captioning scheme: Tier 1 = one high-level video summary (1–2 sentences, no timestamps). Tier 2 = action-level segments with start/end timestamps and clear verb + object labels (these will be pre-annotated and need improvement). Tier 3 = trajectory-level annotations created from scratch: sub-second, body-part-level motion descriptions (may overlap across limbs), grounded only in what is visible (no intent/guesses). Your work will be evaluated on description accuracy, completeness, and timestamp precision. Required Locations: Global - Any Location Required English Level: Fluent Other Qualifications & Requirements (for screening) Confirm C1 English proficiency or higher (comfortable writing precise, natural descriptions). Prior experience with video annotation (action segmentation / temporal labels). Experience with timestamping actions with tight start/end alignment (sub-second precision preferred). Ability to write atomic, observable action labels in verb + object format (e.g., “grasp cup,” “place lid”). Ability to generate trajectory/body-part-level motion descriptions (e.g., left hand/right hand/torso) with < 1 second segments and occasional overlaps. Familiarity with SuperAnnotate (or equivalent annotation tools) and ability to ramp quickly. Proven quality performance on similar projects (targeting 95%+ accuracy and low rework). Availability to support a pilot in February with potential to scale (1000 hours → up to 10,000 hours). Comfort working from customer-provided guidelines and passing a short qualification check before starting.
Total Budget: $9,600Pay Per HourIntermediate1-3 monthsIndependent Labelers Only
Total Budget
$9,600
Pay per Label
$8/hr
Time Requirement
20+ hrs/week
Duration
1-3 months
Description of dataset
Egocentric videos with multi-tier captions
Software
Other
Hiring Type
Independent Labelers Only
Required Location
India, Philippines
Workload / Schedule
Must be able to start Feb 3 and commit 3 hours per day minimum from February 3 - 7
Software
Other
Data Type
VideoLabel Types
Text Generation
Action Recognition
Tracking
Subject Matter / Industry
Human actions and motion trajectories in video
Language
English
Job Type
Managed Service by OpenTrain
Share this project
Share link
trainai.co/project