Principal AI Product Engineer

About the role

A leading organisation in Abu Dhabi that builds and deploys advanced AI products at scale is looking for a Principal AI Product Engineer. This is a senior individual contributor role combining deep AI engineering expertise with strong product judgement, setting the technical standard for how production AI products are designed, built and delivered. The salary range is AED 60-70K.

Key responsibilities

• Own the end-to-end architecture and delivery of production AI products, from AI backends and APIs to user-facing applications.
• Design and establish scalable RAG architectures and Agentic AI workflows, including tool use, planning and human-in-the-loop patterns.
• Define AI-native engineering practices, including coding agents, reusable components, development frameworks and evaluation tooling.
• Build and guide production backend services in Python and modern API technologies.
• Develop reliable AI applications handling streaming, failures and latency.
• Establish evaluation practices covering grounding, retrieval quality, regression detection and production performance.
• Define reusable technical standards, reference architectures and components across AI products.
• Provide technical leadership through architecture reviews, engineering standards and mentorship while staying hands-on with delivery.

Required profile

• Staff, Principal or equivalent senior IC level experience, with a track record of shipping production systems used by real users.
• Deep experience with RAG, retrieval pipelines, Agentic AI, agent orchestration and AI backend services.
• Experience owning AI products end-to-end across backend services, APIs and user interfaces.
• Hands-on experience with modern Agentic AI and LLM frameworks such as LangChain or LangGraph.
• Experience designing reusable engineering standards, tooling and architectures for wider engineering teams.

Required skills

• Strong Python engineering skills and LLM-powered application development.
• Understanding of LLM behaviour and failure modes, including hallucination, context degradation and latency.
• Modern stack: React, APIs, PostgreSQL, Docker, cloud environments and CI/CD.
• AI evaluation frameworks and AI coding agents.