llm orchestration

Turn fragile one-off CLI commands into reliable, shareable Bash workflows with prompts, safe defaults, and previews. This guide shows how to wrap tasks, speed selection with fzf, add guardrails with dialog/whiptail, and drive choices from JSON via jq, with install steps for apt/dnf/zypper and ready-to-use scripts (backup, git branch, env runs, log tails), plus tips for organizing, versioning, and real-world automations to reduce errors and speed onboarding.

Practical, bash-first guide to running multiple LLMs concurrently on one Linux box for specialization, A/B tests, and full GPU use. Covers Ollama, llama.cpp, and vLLM setups; per-GPU/CPU pinning, memory caps, and MIG; apt/dnf/zypper installs; starting/stopping via tmux, systemd, or Docker; Nginx routing behind one URL; and monitoring/tuning—so you can reliably serve, isolate, and compare models in parallel.