The "day 2" discipline — evaluating, deploying, monitoring, upgrading, and governing LLM applications in production. Everything that happens after "it works on my laptop."
LLMOps is the operational discipline — it runs on the AI Platform, operates the Agents, and integrates into the AI-SDLC.
Target audience: ML engineers, SREs, DevOps engineers, and engineering managers running AI in production.
| LLMOps Module | Other Course | Relationship |
|---|---|---|
| M04–M07 Eval & Testing | CE M29 Context A/B Testing | LLMOps automates evals at scale; CE designs what to evaluate |
| M05 LLM-as-Judge | Platform M10 Prompt Eval | Same technique, LLMOps focuses on production automation |
| M08–M11 Deployment | Platform M11, AI-SDLC CI/CD | LLMOps deploys the full app; Platform deploys prompts |
| M12–M15 Monitoring | Platform M28 Observability | App-level monitoring vs Platform's infra-level |
| M13 Quality Drift | CE M14 Context Decay | Same concept — LLMOps detects it; CE prevents it |
| M16–M19 Feedback Loops | Agent M15 HITL | LLMOps systematizes feedback; Agents implement HITL flows |
| M20–M23 Model Lifecycle | Platform M05 Multi-Provider | LLMOps manages upgrades; Platform routes across providers |
| M24–M27 Reliability | Platform M07, M29 | App-level reliability vs Platform's infra-level |
| M28–M30 Governance | Platform M16–M19 Cost | App-level optimization vs org-level budgets |
| M31 Capstone | All capstones | LLMOps capstone operates the prior-authorization pipeline others build |