senior-prompt-engineer

Prompt optimization involves restructuring queries, adding context constraints, and applying chain-of-thought patterns to improve LLM accuracy. This skill provides techniques for Claude, GPT-4, and other production models to minimize hallucinations and ensure consistent outputs across deployments.

RAG evaluation assesses retrieval accuracy and response quality in Retrieval-Augmented Generation pipelines. This skill teaches methods to fine-tune retrieval components, validate source relevance, and measure end-to-end system performance before production deployment.

Agent orchestration structures multi-step workflows where LLMs coordinate tool calls and decisions. This skill covers agent design patterns, state management, fallback strategies, and deployment integration with Python tooling for scalable, observable AI products.

Yes. This skill applies production-grade prompt design, few-shot configuration, and performance tuning directly to Claude, GPT-4, and other commercial LLMs, with built-in monitoring, security, and scalability for enterprise deployments.

You need access to an LLM API, baseline performance metrics, and structured test datasets. This skill assumes Python environment familiarity and requires defined success criteria—response latency, accuracy targets, or cost constraints—before optimization begins.

Structured outputs ensure agents parse responses correctly, reducing downstream errors and enabling reliable tool integration. This skill teaches schema design and validation techniques to enforce predictable outputs across multi-agent systems.