Senior AI Engineer

#### About Homevy Homevy is a holiday homes company in Dubai. Owners sign with us, we manage their property end to end, and we earn a commission on guest bookings. We lead with hospitality and run on technology. Our team covers guest support, housekeeping, maintenance, design and owner onboarding. #### What We're Building We're building the operating system for our property management company, from scratch. It starts with our operations dashboard and grows into a full property management platform: bookings, guests, housekeeping, maintenance and owner reporting in one system we control. AI runs through all of it. Agents answer guests, catch problems before they escalate and coordinate our field team, working from live property and booking data. This is real work with real users from day one. Our guests, owners and operations team will use what you build every day. You will have direct access to the founders, no layers of approval, and the freedom to make the architecture decisions. We need someone who understands AI deeply and can turn that knowledge into reliable software. You should be comfortable designing the architecture, writing the code, connecting systems and improving how they perform in real use. The work includes building strong RAG systems, AI agents and automated workflows that can find the right information, take appropriate actions and involve a person when needed. #### What you’ll do - Own the design and development of Homevy’s AI and automation systems. - Build RAG systems that help AI find and use accurate information from our documents, property records and business systems. - Develop the full process for collecting, preparing, organising and retrieving information, keeping it current as things change. - Build AI agents that can use tools, work through multi-step tasks and interact with our software. - Create workflows that connect different systems, respond to events and recover safely when something fails. - Build capabilities such as guest communication, issue escalation and coordination across operations, while identifying other useful ways to apply AI. - Set clear limits on what AI can do independently and when it must ask for approval or hand over to a person. - Test the quality of AI responses and actions, investigate failures and improve accuracy over time. - Monitor performance, reliability and running costs, and make sensible trade-offs between them. - Protect guest and business information through proper access controls, safe data handling and clear records of automated actions. - Work with our product, engineering and operations teams to turn business needs into systems that work well in practice. #### What we’re looking for - Experience building and running AI systems that real people or businesses use. - A deep understanding of large language models, how they work, where they fail and how to design around those limits. - Strong experience with RAG, including preparing data, embeddings, search, reranking and measuring whether the system retrieves the right information. - Experience building AI agents that use tools and APIs, keep track of tasks and handle errors reliably. - The ability to design complex automations with multiple steps, dependencies, retries and human approvals. - Strong coding skills and good software engineering practices, including testing, documentation and deployment. - The ability to identify why an AI system is failing, whether the problem is the data, retrieval, model, workflow or integration. - Good judgment about which problems need AI and which are better solved with straightforward software. - The ability to work independently, take ownership and explain technical decisions clearly. #### A plus - Backend development and database design experience. - Experience connecting AI to customer support, hospitality, property management, or other business software. - Experience working with both hosted and open-source models, and an understanding of when fine-tuning is useful. #### **When applying** Share examples of AI systems you’ve built. Explain what they do, what you personally worked on, how you tested them and a difficult problem you solved. Links to code, demos, or technical write-ups are welcome.