Artificial Intelligence Engineer

About Property Dynamix

Property Dynamix is a real estate operating system focused on developers in Dubai and Abu Dhabi. It transforms registered DLD and ADREC property transactions into reliable decision-making tools such as unit pricing, pricing simulations, payment plans, land feasibility studies, new launch tracking, promotion monitoring, and comprehensive monthly market reports. Unlike most market platforms that merely sell data, Property Dynamix develops applications that provide actionable insights. Their AI-powered system conducts research on comparable properties, checks pricing accuracy, and performs self-validation before presenting data to users. The AI analyst MOIRA addresses market questions in simple language, and their connector, Majlis, makes market data accessible via Claude and ChatGPT.

Role Overview

This full-time position is a Hybrid/Remote Artificial Intelligence Engineer role based in Dubai. The role centers on building agent systems that translate market data into actionable answers and establishing safeguards to maintain data accuracy and reliability. You will not focus on training neural networks from scratch but will develop multi-agent workflows and orchestration using Google’s Agent Development Kit and other AI tools.

Key Responsibilities

• Design and implement multi-agent workflows with Google’s Agent Development Kit using Python, where each agent performs a specific task and results are combined into cohesive responses.
• Develop tool calling mechanisms with Gemini, managing multi-step tool interactions, structured outputs, and response schemas.
• Write and optimize BigQuery SQL queries for large real estate transaction datasets, transforming query results into tools accessed by agents.
• Create guardrails to prevent AI-generated false data by implementing grounding checks, claim verification, and output rejection gate mechanisms.
• Develop evaluation frameworks and golden test datasets to quantitatively measure improvements to the AI analyst MOIRA’s performance.
• Manage AI model costs and latency by selecting appropriate models per task, routing calls efficiently, and monitoring spending.
• Deploy production systems on Google Cloud platforms including Cloud Run jobs, Firebase Cloud Functions, and scheduled data pipelines.
• Enhance Majlis, the MCP connector integrating Claude and ChatGPT with market data.

Required Qualifications

• Proficiency in Python programming and practical experience with TypeScript.
• Proven hands-on experience developing Large Language Model (LLM) applications in production environments, including tool calling, agent systems, structured output formatting, and prompt design.
• Strong SQL skills, ideally with BigQuery, applied on extensive datasets.
• Experience designing and implementing evaluation metrics, test suites, or scorecards for LLM output quality assessment and iterative improvements.
• Comfortable with deploying and troubleshooting applications on cloud platforms, preferably Google Cloud.
• Detail-oriented with an ability to self-validate work, analyze failures comprehensively, and explain root causes rather than superficial fixes.

Preferred Skills

• Familiarity with Google Agent Development Kit (ADK), LangGraph, or similar agent frameworks.
• Experience with Model Context Protocol (MCP) servers.
• Hands-on knowledge of Firebase services such as Firestore, Authentication, and Cloud Functions.
• Background in real estate, pricing analysis, or financial data.
• Understanding of the Dubai or Abu Dhabi property markets.

Additional Information

• Academic background in Computer Science, Artificial Intelligence, Data Science, or related disciplines, or equivalent professional experience delivering functional systems.
• Salary range is between 20,000 AED and 25,000 AED per month.
• Priority consideration for candidates with experience in UAE real estate pricing and research.
• The position supports hybrid and remote working arrangements.