Custom containers

Bring any deterministic inference workload. Ship a CoreML or MLX-compiled container; we shard, dispatch, and settle.

cc containers push ./my-model.cc
cc jobs run my-model --input data.jsonl