Ai Prompt‑engineering specialist

**AI PROMPT‑ENGINEERING SPECIALIST**
Crafting, evaluating, and hardening the prompts that keep the governance, risk and compliance (GRC) platform's large‑language‑model services accurate, secure, and regulator‑ready.

**Location**: Can work remote plus visit the office in Alexandria, Egypt (San Stefano) for team meetings.

**Job Type**: Full‑time, contractor
**Reports to**: AI Systems Architect

PROMPT ENGINEERING

Do you relish the puzzle of squeezing maximum signal—and zero hallucination—out of an LLM? Can you translate security questionnaires, SIEM anomalies, and GRC policy clauses into retrieval‑augmented prompts that withstand red‑team probing? As an AI Prompt‑Engineering Specialist you will design and iteratively refine the system, user, and chain‑of‑thought prompts that power every inference endpoint of the GRC platform—from auto‑answering ISO‑27001 customer audits to recommending CVE vulnerability remediations.

BACKGROUND

Allendevaux & Company is fusing global regulatory expertise with cutting‑edge AI to build a multi‑module GRC platform. You will collaborate with ML/LLM Engineers, MLOps Engineers, and UX researchers in Alexandria to ship low‑latency, compliant LLM services based on models such as Llama 3 and Mixtral 8×7B.

EXPERIENCE

PERSONAL CHARACTERISTICS & SKILLS
- Clear, precise English; able to write prompts and explanations that non‑experts can trust.
- Methodical experimenter who validates prompt changes with quantitative metrics.
- Security‑aware, understanding jailbreak vectors and alignment guard‑rails.
- Collaborative—works hand‑in‑glove with ML, DevOps and UX peers.
- Curious, keeping pace with the fast‑moving LLM and prompt‑engineering research canon.

RESPONSIBILITIES
- Designing system, user, and function‑calling prompts that drive accurate, context‑rich answers.
- Building prompt‑evaluation suites (A/B, canary, red‑team) to measure latency, accuracy, hallucination, bias and toxicity.
- Working with ML Engineers to integrate prompts into RAG pipelines and tool‑calling chains.
- Developing prompt templates for multilingual support, policy‑citation mode, and regulator‑safe explanations.
- Documenting prompt versions, rationale, and experimental results for auditability under ISO 42001 and EU AI Act guidance.
- Collaborating with UX researchers to test answer clarity and user trust.
- Monitoring production metrics and refining prompts to reduce cost and improve relevance.

MINIMUM QUALIFICATIONS
- Bachelor’s degree in Computer Science, Computational Linguistics or related field; **master’s degree highly preferred**.
- 2 + years working with transformer‑based LLMs or conversational AI systems.
- Proficiency in Python plus one LLM framework (LangChain, LlamaIndex, or similar).
- Experience with prompt‑evaluation metrics and retrieval techniques.

PREFERRED QUALIFICATIONS
- Exposure to GRC, cybersecurity, or regulatory language corpora.
- Familiarity with bias/fairness testing tools (e.g., Aequitas, Fairbench).
- Knowledge of vector‑database tuning and embedding‑model selection.
- Publications, blog posts, or competition entries (Kaggle, HackAPrompt) in prompt design or NLP.