機器學習
LLMOps Skill
Use this skill for LLMOps tasks: prompt engineering, model versioning (MLflow/DVC), model serving (FastAPI/Triton), feature store patterns, A/B testing for models, LLM observability (Langfuse/Phoenix/Helicone), guardrails and safety, fine-tuning and RLHF guidance, LLM evaluation frameworks, and AI cost optimisation. Triggers: "prompt engineering", "fine-tuning", "RLHF", "model versioning", "MLflow", "model serving", "A/B test model", "LLM observability", "Langfuse", "guardrails", "LLM evaluation", "evals", "AI cost", "token optimisation", "model deployment", "inference", "LLMOps".