AI/ML Engineer — Point Cloud to Floorplan Automation

About This Role

About Robotic Imaging

Robotic Imaging – FZCO is a Dubai-based technology company specialising in 3D laser scanning, AI development, and design services. Scanning buildings with LiDAR is our core business — we capture sites every week and convert the resulting point clouds into Revit and CAD deliverables for architects, engineers, and contractors worldwide. We operate from IFZA Business Park in Dubai Silicon Oasis, with collaborating teams in the United States, India, and Pakistan.

The Opportunity

We are hiring an AI/ML Engineer to lead the development of software that automatically generates accurate 2D floorplans — and ideally 3D building elements — directly from raw point cloud data. Today, this conversion is largely manual, with modelers spending days tracing walls, openings, and fixtures. Solving this is the single highest-leverage problem in our business, and the successful candidate will own it end-to-end, reporting directly to the CEO.

Key Responsibilities

• Design, prototype, and productionise a pipeline that converts point cloud data (.e57, .las, .rcp, .pts) into usable floorplan outputs (DWG, DXF, IFC).
• Evaluate and combine approaches across classical geometry (RANSAC, Hough transforms, slice-based extraction) and modern deep learning (PointNet++, KPConv, RandLA-Net, Point Transformer).
• Explore the use of vision and multimodal LLMs fine-tuned for AEC-specific tasks where appropriate.
• Build training and evaluation pipelines using our in-house library of real-world point clouds and validated ground-truth BIM models.
• Benchmark model accuracy against manual deliverables and iterate toward production-grade reliability.
• Collaborate daily with our scanning and modeling teams to refine requirements and validate outputs.
• Document architecture, methodology, and results clearly for both technical and non-technical stakeholders.

Required Qualifications

• 4+ years of hands-on experience in computer vision, 3D deep learning, or geometric processing.
• Strong proficiency in Python with NumPy, PyTorch or TensorFlow, Open3D, and PCL.
• Demonstrated production experience with 3D data (point clouds, meshes, voxel grids) — not just tutorials or coursework.
• Proven ability to own a model end-to-end: data preparation, training, evaluation, and deployment.
• Comfortable working on-site in Dubai and collaborating across multiple time zones.

Preferred Qualifications (Major Plus)

• Prior experience with AEC formats and tools (IFC, DWG, Revit API) and scan-to-BIM workflows.
• Background in photogrammetry or reality capture.
• Published research or open-source contributions in 3D vision or geometric deep learning.

Why This Role Stands Out

• Direct reporting line to the CEO — fast decisions, no bureaucracy.
• Access to a large, continuously growing library of real-world LiDAR point clouds paired with validated ground-truth models. You will not be hunting for training data.
• You will be building the product, not a feature inside someone else's product.
• Open to either full-time employment or a senior contractor engagement for the right candidate.

Compensation & Benefits (Full-Time)

• AED 24,000 – 28,000 per month, based on experience.
• UAE employment visa and Emirates ID, fully sponsored by the Company.
• DHA-compliant health insurance from day one.
• 30 days annual leave plus all UAE public holidays.
• End-of-service gratuity in accordance with UAE Federal Decree-Law No. 33 of 2021.