AI Prompt Specialist
Job summary
We are seeking an AI prompt Engineer who will design, test, and refine text or image inputs for generative AI models to produce accurate, safe, and context-aware outputs. A person who will bridge human intent with machine intelligence by optimizing prompt structures to improve AI relevance, reduce bias, and ensure high-quality.
Job descriptions & requirements
Responsibilities:
- Create and refine prompts for Large Language Models (LLMs) to maximize accuracy and creativity.
- Analyze AI outputs to identify limitations, reduce "hallucinations," and improve consistency.
- Develop and manage a repository of effective prompt templates for team-wide use.
- Work with developers, product managers, and data scientists to align AI behaviors with business needs.
- Help refine models to adopt specific brand voices or specialized, domain-specific knowledge.
- Develop prompt libraries, documentation, and best practices. Collaborate with product, marketing, engineering, and data teams.
Requirements:
- 5 years of proven experience working with generative AI tools (e.g., ChatGPT, Gemini, Claude, Midjourney).
- Degrees in Computer Science, Linguistics, Cognitive Psychology, or Data Science are highly valued, though not always mandatory.
- Certificates in AI, NLP, or specific prompt engineering courses can strengthen a resume.
- A deep understanding of how LLMs, such as GPT-4, Claude, and Gemini, work is essential. This includes tokenization, context windows, and parameters like "temperature".
- Proficiency in Python is often required for scripting, automation, and interacting with AI APIs. Familiarity with libraries like TensorFlow or PyTorch is beneficial.
- The ability to analyze AI outputs for accuracy, bias, and relevance to refine prompts is necessary.
- Excellent technical documentation and communication skills.
Remuneration: NGN 500,000
Important safety tips
- Do not make any payment without confirming with the Jobberman Customer Support Team.
- If you think this advert is not genuine, please report it via the Report Job link below.