ETIC Capture

Real-world video data for embodied AI training. Captured on request.

Robotics teams define the task. I capture it on real jobsites, with real tools and real workers, then deliver organized, labeled footage ready for training pipelines.

Scope

Requested tasks

Teams specify the action, setting, objects, camera angle, repetition count, and any capture constraints. I confirm feasibility before committing to a shoot.

Blue-collar workflows

Capture areas focus on blue-collar and jobsite work: material handling, tool use, packing, inspection, repairs, cleaning, sorting, shop tasks, and site movement. These are environments I have direct access to and working knowledge of.

Structured delivery

Footage is organized by task, environment, object, worker, attempt, outcome, and timestamp, with notes and metadata structured for downstream use. Each delivery includes a manifest and capture log.

Legally safe use

Worker consent, location permissions, and usage terms are handled in writing before any recording begins.

Process

  1. 1

    Brief

    You send the task, requirements, constraints, and target format.

  2. 2

    Capture

    Once feasibility is confirmed, I record real execution with clean framing.

  3. 3

    Structure

    Files are labeled, tagged, versioned, and packaged with rights documentation.

  4. 4

    Deliver

    Final dataset arrives in your specified format with a full manifest.

Contact

If real-world physical data would support your team's training work, get in touch.