Dwelleo is a Saudi AI-powered PropTech platform redefining how real estate decisions are made. From buying and renting to selling and investing, we combine verified data, intelligent insights, and complete transparency empowering users to understand the market clearly and make confident decisions at every step.
About Dwelleo
Dwelleo is an AI-powered real estate marketplace transforming how people search, buy, sell, and rent properties across Saudi Arabia.
The platform combines machine learning, intelligent discovery tools, and data-driven insights to connect buyers, renters, brokers, and developers through a seamless, scalable digital experience.
At its core, Dwelleo embeds AI directly into the product powering pricing, recommendations, search, and decision-making across the entire property journey.
About The Role
You'll work across the full AI surface area of a live product. Your work will span classical ML (pricing, forecasting) and modern LLM systems (agents, RAG, voice), with real engineering expectations around reliability, observability, and production quality. Research here is always in service of shipping experimentation exists to accelerate delivery, not replace it.
You'll work within a dual-track Agile structure: a
Research sprint for experimentation and model development, and an
Engineering sprint for deployment, integration, and reliability work. You'll contribute to both.
What You'll Do
- Ship predictive models for property price estimation, rent forecasting, and ROI analysis from feature engineering through production deployment on AWS ECS with canary release practices
- Build and maintain LLM-powered multi-agent systems, including deterministic state machines, scoring engines, guardrail layers, and escalation logic that runs reliably at scale
- Develop and improve RAG pipelines with vector-based retrieval (FAISS, pgvector) for property search and AI-assisted advisory features
- Expose AI capabilities as FastAPI microservices with autoscaling, versioning, and production-grade reliability
- Integrate and improve STT/TTS speech pipelines connected to telephony infrastructure Arabic dialect support is a meaningful plus
- Own observability across the stack: drift detection, latency monitoring, alerting, and model performance dashboards
- Collaborate closely with product and engineering across both sprint tracks, translating research outcomes into shippable systems
Required
What we're looking for
- 36 years of hands-on experience in ML engineering, AI engineering, or a closely related role with at least some of that time shipping to production
- Strong Python skills with a track record of production deployment (not just notebooks)
- Solid ML fundamentals: supervised learning, gradient boosting (XGBoost / LightGBM), feature engineering, model evaluation, and retraining workflows
- Hands-on experience with LLMs (OpenAI, Anthropic, or open-source models) and practical RAG pipeline implementation
- Experience building and deploying ML microservices using FastAPI
- Working knowledge of CI/CD for ML workloads you've set up or maintained pipelines using tools like GitHub Actions, MLflow, or equivalent
- Comfortable working in English; Arabic is not required but welcome
Nice To Have
- STT/TTS integration or voice pipeline engineering Arabic dialect experience is a strong plus
- AWS hands-on experience: ECS, S3, RDS
- Familiarity with geospatial data and similarity search (Ball Tree, FAISS, pgvector)
- Background in lead scoring, content moderation, or recommendation systems
- Experience with MLflow, experiment tracking, or model registry workflows
- Real estate or PropTech domain exposure
What We Offer
- Location: Fully remote
- Team: You'll report directly to the AI Team Lead
- Direct ownership of production AI systems with measurable user and business impact
- A technically ambitious environment working at the intersection of AI and one of the region's most active real estate markets
