senior-prompt-engineer

Design robust prompt patterns by applying established techniques like Zero-shot, Few-shot, ReAct, and Chain-of-Thought to diverse tasks. This approach structures LLM workflows to improve reliability and efficiency across prompt optimization and agent architectures.

Evaluate LLM prompt performance by creating structured evaluation frameworks and instruments. These frameworks compare prompts and agent behaviors systematically, providing measurable guardrails and generating detailed evaluation reports.

Architect agent workflows by applying prompt-driven patterns like ReAct, Plan-Execute, Tool Use, and Multi-Agent. These patterns establish measurable guardrails and structured outputs, enabling robust production-grade LLM workflows.

Structured output formats like Mermaid and ASCII diagrams are supported for production-grade prompt engineering. These formats help visualize agent architectures and workflows, alongside cost estimates and evaluation reports.

Structured evaluation frameworks are needed when comparing prompt behaviors or optimizing LLM workflows for production. They provide validation tooling and measurable guardrails to ensure reliability across diverse prompt engineering patterns.

Measurable guardrails and token-usage analysis are needed to control costs and ensure reliability in LLM workflows. They provide structured cost estimates and validation tooling to support production-grade prompt engineering.