Ways to work together

Pick the engagement that fits the problem.

Six models cover everything from a licensed pack to a staffed, continuous operation. Each lists what it fits, the scope, its boundary and how it is billed.

E1Non-custom

Standard Dataset License

You need proven data now and can work from licensable inventory.

Scope · Select fixed tasks, fields and versions from licensable inventory; delivered for agreed purposes and license scope.

Boundary · Licensable scope, volume and update frequency are governed by the specific inventory.

Dataset / task pack / license scope

E2Filtered + processed

Semi-Custom Data Pack

Existing data fits, but you need it filtered and shaped to your pipeline.

Scope · Filter available data by scene, object, language or segment; add specified labels, splits or format conversion.

Boundary · Built on available inventory; new capture is out of scope for this model.

License fee + filtering / processing fee

E3New capture

Fully Custom Collection

You need new tasks, scenes, devices or protocols captured from scratch.

Scope · Capture new tasks, scenes, devices and protocols; includes design, deployment, capture, annotation and acceptance.

Boundary · Configuration and acceptance defined per project before batch production.

Project / milestone; qualified participants, hours, clips or episodes

E4Exclusivity

Dedicated or Exclusive Program

You require exclusivity over a task, domain or dataset.

Scope · Define exclusivity by task, domain, term or customer; set reuse boundaries for original and derived data.

Boundary · Reuse of original and derived data is bounded by the agreed exclusivity scope.

Dedicated project fee + exclusivity-scope fee

E5Ongoing

Continuous Data Service

You need a recurring pipeline that keeps improving after deployment.

Scope · Periodic new collection, log feedback, failure recapture, updates and continuous evaluation.

Boundary · Reserved capacity and delivered volume are agreed per period.

Monthly service / reserved capacity + delivered volume

E6Staffed

Managed Team & Experts

You want a staffed team operating your data workflow under your direction.

Scope · Dedicated or shared operators, annotators, experts and project management, staffed by task complexity.

Boundary · Team composition and permission scope are specified per project.

Team-month / expert hours / accepted output

The project path

From requirements to ongoing operation.

  1. 01Requirements & sample review
  2. 02Specification & quotation
  3. 03Small pilot
  4. 04Adjustment based on results
  5. 05Batch delivery & acceptance
  6. 06Updates or ongoing operation

What we scope with you

Six dimensions that turn a goal into a specification.

01

Model & goal

Training / fine-tuning / evaluation stage, task list, model interface, application and success criteria.

02

Source & modality

New or existing; real robot / human demonstration / simulation; images, video, audio, state/action and sensors.

03

Scale & coverage

Independent participants, tasks, objects and scenes; valid hours, clips or episodes; long-tail and failure ratio.

04

Technical & labels

Resolution, frequency, synchronization, coordinates, control definitions; label ontology, precision, language, fields and format.

05

Usage & delivery

Permitted use, exclusivity scope, delivery location/method, schedule, batches, update frequency and versioning.

06

Acceptance & settlement

Valid-sample definition, sampling plan, quality thresholds, rework/recapture rules, acceptance cycle and billing unit.

Start an inquiry

Tell us the goal. We will propose a pilot.

Step one takes a minute. Add detailed scoping if you already know your tasks, scale and formats — it makes the first quotation sharper. Product codes and the engagement model you arrived with are carried in automatically.

Loading form…