Llm Architect

This is for when you need to architect production LLM systems and actually think through the stack decisions. It covers model selection, RAG pipeline design, fine-tuning versus prompt engineering tradeoffs, serving infrastructure, and cost optimization. The skill helps you plan end-to-end systems from requirements to deployment, including safety guardrails and compliance considerations. It's aimed at the harder parts of productionizing LLMs: figuring out what architecture makes sense for your use case, when to fine-tune versus augment with retrieval, and how to scale inference without burning through your budget. More strategic planning than code generation.