AI Engineer

UMATR; Dubai; On-site; AED 335,000 - 391,000; Full time

Job summary

UMATR in Dubai is seeking a hands-on AI/software engineer to join our production AI platform. You will build AI systems with Python, develop LLM-enabled workflows, and contribute to RAG and retrieval infrastructure.

Work closely with engineers to translate complex operational challenges into reliable AI solutions, from design through deployment. On-site role in Dubai with competitive AED 30,000–35,000 monthly.

Qualifications

• 4+ years of experience developing AI applications or production AI systems.
• Strong hands-on Python and backend development.
• Experience building AI-powered applications with LLMs, RAG or tool calling.
• Familiarity with REST APIs and integrating AI into existing apps.
• Familiarity with vector databases and semantic search.
• Able to collaborate with engineers and review code for production systems.
• Proactive problem-solver with willingness to learn new AI tech.

Responsibilities

• Build and maintain production AI systems using Python.
• Develop AI agents and automated workflows with LLMs.
• Design and improve RAG/GraphRAG with vector databases.
• Build AI evaluation frameworks for retrieval quality and agent performance.
• Benchmark open-source LLMs and contribute to self-hosted infrastructure.
• Improve system reliability via monitoring and automated testing.
• Collaborate with engineering and business teams to translate challenges.
• Own development from design to deployment and ongoing improvement.

Skills

Python

Backend development

LLMs

REST APIs

Vector databases

AI systems

Team collaboration

Production AI

Tools

Docker

Kubernetes

Job description

On-site in Dubai

AED 30,000 - 35,000

About the Company

We're partnering with a forward-thinking organisation that's going beyond simply adopting AI, they're building it into the way their entire business operates.

With AI systems already running in production, they're investing in the next generation of intelligent technology, from autonomous AI agents and advanced retrieval systems to self-hosted LLMs.

And this is where things get exciting.

Responsibilities

• Build and maintain production AI systems using Python, developing backend services and intelligent applications.
• Develop AI agents and automated workflows, including LLM integrations, tool calling, routing and orchestration.
• Design and improve RAG and GraphRAG systems, working with knowledge graphs, vector databases and semantic search.
• Build AI evaluation frameworks to measure retrieval quality, agent performance and reliability.
• Benchmark open-source LLMs and contribute to the transition towards self-hosted model infrastructure.
• Improve system reliability through monitoring, observability, automated testing and production support.
• Collaborate with engineering and business teams to translate operational challenges into practical AI solutions.
• Take ownership of development from initial design through to deployment and ongoing improvement.

Qualifications

• 4+ years of experience developing AI applications or working in AI/software engineering roles involving production AI systems.
• Strong hands-on experience with Python and backend development.
• Hands-on experience building AI-powered applications using LLMs, with exposure to technologies such as RAG, AI agents or tool calling.
• Understanding of REST APIs and integrating AI capabilities into existing applications.
• Familiarity with vector databases, semantic search or retrieval-based systems.
• Comfortable working with other engineers, reviewing code and contributing to production applications.
• A proactive approach to problem-solving and an interest in learning new AI technologies.

Preferred Skills

Experience in any of the following would be beneficial, but you don't need to have worked with all of them:

• GraphRAG, knowledge graphs or ontologies.
• LLM evaluation, retrieval metrics and regression testing.
• PyTorch, Hugging Face, Ollama or vLLM.
• Qdrant, Milvus or other vector databases.
• Docker, Kubernetes or model-serving infrastructure.