Oldest Kid — Infrastructure
Dirty face, overalls, toolbelt. Kevin is the oldest kid — the one who keeps the local infrastructure humming while everyone else gets the glory. He manages your LLM, keeps Ollama running, handles model downloads, and makes sure the family has the compute they need. Named after his uncle, though nobody talks about why.
Kevin keeps your local LLM running. Model pulls, health checks, context window management. If Ollama needs attention, Kevin handles it before you notice.
Makes sure every family member can reach the services they need. Port management, process lifecycle, dependency ordering — the plumbing nobody thanks you for.
Tracks which models are downloaded, which are active, and which need updating. Vera asks for an embedding model, Kevin makes sure it's there.
Your laptop has finite RAM and GPU. Kevin decides who gets what, when, and makes sure the family doesn't starve your other applications.
Monitors local service health. If Ollama crashes at 3 AM during one of Norm's research runs, Kevin restarts it. Automatically. Quietly.
Kevin detects what's available on your machine and adapts. GPU present? Use it. CPU only? Adjust model selection. No configuration files to edit.
Kevin the kid manages your local machine — Ollama, models, services. He's a teenager with a toolbelt, not a datacenter. His uncle handles the serious enterprise work overseas. Different Kevin, different scale, different story entirely.