Synthetic dataset generation

Generate, filter, deduplicate, review, and publish synthetic records.

define schema and policy→generate batch Agenerate batch Bgenerate batch N→filter→deduplicate→quality review→publish dataset

Synthetic dataset generation

Activity responsibilities

Activity	Responsibility
define-schema-and-policy	Execute the define schema and policy stage and publish its declared outputs for downstream activities.
generate-batch-a	Execute the generate batch A stage and publish its declared outputs for downstream activities.
generate-batch-b	Execute the generate batch B stage and publish its declared outputs for downstream activities.
generate-batch-n	Execute the generate batch N stage and publish its declared outputs for downstream activities.
filter	Execute the filter stage and publish its declared outputs for downstream activities.
deduplicate	Execute the deduplicate stage and publish its declared outputs for downstream activities.
quality-review	Execute the quality review stage and publish its declared outputs for downstream activities.
publish-dataset	Execute the publish dataset stage and publish its declared outputs for downstream activities.

Inputs

• Schema and policy
• Source examples
• Generator model

Outputs

• Raw batches
• Validated dataset
• Quality report

Provenance to preserve

• Prompt and model
• Filter reasons
• Review decisions

Runnable example

This is the complete checked-in bundle for this pattern. Download the environment, scope, topology, workflow, input, container recipe, runner, and validation contract from this page before executing it.

Read setup instructions ↓Environment YAML ↓Execution scope YAML ↓Network topology YAML ↓Workflow YAML ↓Input data ↓Container recipe ↓Run the workflow ↓Validate outputs ↓Expected output manifest ↓

Verified local execution

These captures and the output manifest were produced by the fixture's local Docker run and validator. They are published with the same bundle as the runnable files.

Workflow evidence

Execution evidence

Output evidence

Open verified output manifest ↓

Execution considerations

Use queued external jobs for rate-limited generation. Stable record identities prevent duplicates after retries.

AkôFlow boundary: the engine schedules, deploys, executes, transfers data, and records evidence. The ML or agent framework remains an implementation choice inside each activity.