HarborExpert sign in

Harbor Expert Network

Robotics RGB-D Capture Specialist

$90–$140/hrpay

Required skills

RGB-D captureROSSensor calibrationManipulation datasets

About Harbor

Harbor connects domain experts to the development of frontier AI models. Real-world expertise is turned into training data, evaluations, and feedback loops that improve how models perform. AI labs and enterprises use Harbor to train models and build reliable agents through advanced evaluations and field capture programs. Experts contribute directly to how AI systems learn, reason, and perform across robotics, computer vision, wearable AI, industrial inspection, healthcare, and agentic systems. Harbor identifies and vets top talent through an AI recruiter, enabling high-quality contributions at scale.

Job description

Job title: Robotics RGB-D Capture Specialist

Job type: Contractor

Location: Hybrid

Job summary: In this role, you will capture and validate RGB-D datasets for robotics and embodied AI programs. Your work shapes how manipulation and navigation models learn from real-world sensor data. No prior AI industry experience is required — hands-on robotics and capture discipline matter most.

Key responsibilities

  1. Plan and execute RGB-D capture sessions for manipulation, navigation, and teleoperation programs.
  2. Calibrate depth cameras, verify time sync across sensors, and document coordinate frames and rig setup.
  3. Review capture quality for occlusions, motion blur, slip events, and safety protocol adherence.
  4. Annotate metadata and flag edge cases that would poison training or sim-to-real pipelines.
  5. Collaborate with Harbor program leads on guideline updates and failure-mode taxonomies.
  6. Deliver consistent, audit-ready capture packages on schedule in lab or field environments.

Required skills and qualifications

  1. Hands-on experience with depth cameras, gripper rigs, or mobile manipulation platforms.
  2. Working knowledge of ROS or equivalent robotics middleware and calibration workflows.
  3. Ability to judge grasp quality, collision risk, and label-worthy vs discard frames.
  4. Strong written communication for session logs, metadata, and QA notes.
  5. Comfort operating under safety and access protocols on factory or lab floors.

Preferred qualifications

  1. Experience with Isaac Sim, teleoperation stations, or industrial pick-and-place lines.
  2. Background in computer vision QA or robotics research publications.