Local Inference Server Pattern

Use a bank of local machines (e.g., Mac Minis) as on-prem inference servers acting as remote endpoints for lightweight open-source models.