Migrate from Azure Batch

Azure Batch (or Azure Machine Learning batch endpoints) maps cleanly to Common Compute for the inference half of your pipeline. The same drop-in shim that works for AWS Batch is on the roadmap; today the path is the native SDK.

Concept mapping

• Pool → none (we own the fleet)
• Job + Task → workload + job
• Application package → workload_id + model_id
• Blob storage input → R2 presigned upload
• Managed identity → API key with `inference` scope
• Batch account → Common Compute account (one per organisation)

Migration code

from commoncompute import Client

cc = Client()
job = cc.jobs.submit(
    workload_id="coreml_embed",
    model_id="bge-large",
    payload={"texts": [...]},
    priority="batch",
    max_spend_usd=50.00,
)