Covers the full production deployment stack for ML models: FastAPI server setup, Docker containerization, Kubernetes orchestration, and monitoring. The real value is in the troubleshooting section, which walks through eight common failure modes like missing health checks causing 503s, OOM kills from missing resource limits, and silent model drift. Includes a practical six-step quickstart and a deployment checklist. If you're moving models from notebooks to production and keep hitting the same infrastructure problems, this gives you working patterns for API endpoints, batch processing, rollback procedures, and CI/CD validation. Less about the ML theory, more about keeping services up and catching issues before users do.
npx skills add https://github.com/secondsky/claude-skills --skill model-deployment