Prompt Engineer Resume: What to Put On It

A prompt engineer's resume needs to demonstrate their evaluation methods, cost and latency constraints, and how their work holds up when the model changes. Simply pasting a great prompt into the resume is not enough, as it only shows what you typed, not what you know. Job titles for prompt engineers can vary, such as AI engineer, LLM application engineer, or ML engineer (GenAI), so a resume should adapt to these different titles.

The resume should showcase an evaluation loop with labeled sets, pass rate, and a way to distinguish real improvement from lucky samples. Production constraints, such as token cost per request, p95 latency, and context limits, are crucial. Robustness requires prompts to survive model version changes, adversarial input, and traffic not resembling test cases.

Engineering around the model involves retrieval, structured output, schema validation, fallbacks, guardrails, and human review routing. The line "Wrote prompts that improved accuracy" is often the downfall of a resume, as it fails to provide concrete evidence. When writing bullet points, follow an engineering structure: what you built, on what scale, with measured results.

Use real numbers from evaluation logs or dashboards rather than making up plausible percentages. Include details about the evaluation set size and shape, baseline and lift, cost and latency, human workload, and durability. In the skills section, name the stack but avoid listing only tool names. Group them to argue for your experience.

If you don't have a prompt engineer job title yet, put LLM work at the top of relevant role bullets, even if it was a small portion of the job. Use a projects section for work outside the job, and honestly name adjacent titles that reflect your experience. Avoid prompt galleries and instead link to a repository with evaluation scripts and results.

A well-formatted resume, with one column of standard text and consistent section headings, is essential.