AI Prompt Engineer Jobs in Brazil (Remote)
Design, test, and evaluate prompts and system instructions to improve large language model outputs for production workflows. You will build prompt libraries, run A/B tests, support RLHF-ready artifacts, and partner with engineering teams to ship reliable prompt templates for tools, agents, and RAG use cases.
Key Responsibilities
- Build reusable prompt libraries, templates, and style guides
- Run prompt evaluation, A/B tests, regression suites, and red-teaming
- Define evaluation criteria, QA rubrics, and scoring guidelines
- Create RLHF-ready artifacts (preference comparisons, prompt-response pairs)
- Analyze failure modes (hallucinations, toxicity, policy violations) and iterate
- Document experiments, metrics, and learnings for repeatable improvements
Required Qualifications
- Production experience in prompt engineering and/or LLM evaluation programs
- Strong writing, structured reasoning, and experiment design skills
- Familiarity with NLP concepts and LLM failure modes
- Ability to write clear annotation guidelines and QA checklists
- Comfort collaborating remotely with cross-functional stakeholders
Preferred
- RLHF / human-in-the-loop evaluation and preference ranking experience
- RAG prompting, tool calling, or agentic workflow experience
- Python/SQL for analysis, telemetry, and experiment tracking
- Content safety, policy, and adversarial prompting exposure
Pay
$30–$50 per hour (base).