Start a program

AuraOne / AI Labs

Choose the operating product for the AI work.

AI Labs is the compatibility route for teams choosing Human Data, Models, or App Data. Models is the primary name for evaluation, training, release, and model operations.

Existing AI Labs routes and terminology remain supported where needed. New evaluation, release, and model-operations journeys use the Models name.

Open ModelsRoute a program

Product paths

Start from the input, work, and decision.

Human Data

Qualified people create, review, and improve the data AI systems use, with training, rights, quality, and delivery evidence attached.

Explore Human Data

Models

Evaluate, compare, improve, approve, release, and monitor models against explicit criteria and retained regressions.

Explore Models

Compute

Search, compare, reserve, and operate GPU capacity across a provider network.

Explore Compute

App Data

Choose a domain application, provide the input, review the work, and receive a usable, evidence-linked output.

Explore App Data

A shared path from work to evidence

The product changes with the object and decision, but the evidence contract remains recognizable.

  1. 01

    Choose the object

    People and data, a model or agent operation, or a domain input.

    Product and source

  2. 02

    Define the criteria

    Rights, quality, evaluation, policy, review, and acceptance are explicit.

    Versioned requirements

  3. 03

    Run and review

    Work, exceptions, failures, reviewer judgment, and ownership remain visible.

    Work and review history

  4. 04

    Approve the output

    Deliver, release, hold, return, or operate with an evidence packet.

    Decision and output record

Choose by the decision the team must make.

Choose by the decision the team must make..
OutcomeWorkEvidenceBoundary
Accept specialist-created dataQualify people, create or review work, adjudicate quality, and package delivery.Qualification, rights, task, rubric, review, and manifest.Use Human Data for people-and-data programs.
Ship, hold, or improve a modelEvaluate candidates, replay failures, compare, approve, release, and monitor.Source runs, comparison, regressions, blockers, approvals, and launch file.Use Models for evaluation, training, release, and model operations.
Release a domain-specific outputValidate a domain input, run the workflow, review evidence, and create a handoff.Input, stages, reviewers, artifact, output, manifest, and readiness.Use App Data when a domain application matches the input and output.