general-prompt-engineer

Prompt optimization involves diagnosing failures, removing contradictions, and producing compact, testable prompts with labeled sections or schemas to ensure clear, consistent, and reliably evaluable AI responses across tasks.

Prompt debugging involves diagnosing structural failures, removing contradictions, and applying validation rules to produce testable prompt drafts with lightweight eval plans to catch hallucinations and measure reliability.

Schema design for agent workflows requires recommending labeled sections, schemas, or staged workflows to enforce machine-validated outputs, safe tool use, and reliable format enforcement.

Generating model-family variants for a system message provides lightweight alternative prompt drafts to measure reliability and maintain portability across different AI model architectures.

Creating eval rubrics provides lightweight evaluation plans and assumptions that deliver a short accuracy checklist to measure reliability and catch hallucinations in AI outputs.

Prompt compression for tool instructions produces compact, testable prompt drafts while retaining structural grounding and format enforcement needed for safe tool use and automation.