GenAI Engineer

About the Job: As a GenAI Engineer, you will leverage the company’s no-code builder to design and build AI agents and GenAI applications. This full-time hybrid role based in Dubai requires an individual proficient in both Arabic and English, with a focus on creating innovative solutions and ensuring the quality of AI outputs.

Responsibilities:

• Solution Design & Build: Utilize the company’s no-code builder to create AI agents and GenAI applications using Application, Automation, Data, and AI Agent modules.
• Prompting & RAG: Engineer prompts, tool/function calling, chunking/embedding, vectorization, reranking, grounding, and guardrails to ensure high-quality outputs in Arabic and English.
• Agent Evaluation: Create golden test sets and utilize agent-eval frameworks to track accuracy.
• LLM Model Training & Fine-Tuning: Conduct training and fine-tuning of models through prompt engineering and other methods in Arabic and English.
• Customer Delivery: Execute discovery processes, define user stories and acceptance criteria, build Proofs of Concept (PoCs) and Minimum Viable Products (MVPs), and iterate rapidly with stakeholders.
• Platform Expertise: Serve as a Subject Matter Expert (SME) on the company platform and deliver structured product feedback to influence the product roadmap.

Requirements:

• Experience: 2-4 years in AI or LLMs, focusing on building real applications or agents.
• Language Proficiency: Proficient in both Arabic and English (read/write/communicate).
• In-depth knowledge on Gen AI: Familiarity with LLMs, teams of agents, MCP/A2A, RAG (chunking, embeddings, reranking, vector databases), grounding/hallucination, AI governance, auditability, evaluation, and Human-In-The-Loop (HITL) processes.
• Model Ops Skills: Experience in prompt engineering, creating golden test sets, agent-eval frameworks, A/B testing, regression analysis, and basic fine-tuning techniques (PEFT/LoRA).
• Integration Skills: Knowledge of REST/GraphQL APIs, OAuth, webhooks; ability to stitch systems into end-to-end flows.