Compute. The compute backbone underneath it all: batch scoring, large simulation runs, and burst capacity, managed.
When a client asks you to score 10,000 documents overnight, ServerCore is the reason your answer is “sure” instead of a quote for a cluster. Jobs are queued, parallelized, checkpointed, and billed by the unit of work, not by the hour you forgot to turn something off.
job = client.groundwater.servercore.batch(
task="radiance.rootcause.score",
inputs=s3_manifest("client-archive-2025"),
priority="overnight"
)
job.on_complete(webhook="https://yourfirm.com/hooks/scoring-done")Back to Groundwater, or the reference overview.