Senior AI - Computer Vision Engineer

The role

You will develop perception and scene-understanding systems for wearable augmented reality technology. This involves building algorithms for 3D environments, spatial intelligence, and object recognition using real-world sensor data.

What you will do

• Develop advanced computer vision algorithms for construction scene understanding and spatial intelligence.
• Design and develop semantic segmentation, object detection and instance segmentation models.
• Build algorithms capable of understanding 3D environments, spatial relationships and physical geometry.
• Develop automated BIM correspondence and object-matching capabilities.
• Develop algorithms using multi-modal sensor data.
• Combine AI, 3D vision, SLAM and BIM to develop geometric and spatial reasoning capabilities.
• Explore emerging scene representation techniques including Gaussian Splatting and NeRFs.
• Improve localisation-assisted perception and robustness across varying lighting, weather and real-world site conditions.
• Design, train and evaluate large-scale deep learning models.
• Develop data pipelines supporting supervised, self-supervised and semi-supervised learning.

Required qualifications

• 5+ years’ industrial experience in Computer Vision
• MSc or PhD in Computer Vision, AI, Computer Science or a related discipline
• Deep expertise in 3D Computer Vision / Spatial AI
• Strong understanding of geometry, spatial relationships and scene understanding
• Experience developing and deploying computer vision and deep learning systems in production
• Strong Python development skills
• Expert knowledge of PyTorch
• Experience with semantic segmentation, object detection, image understanding and modern computer vision architectures
• Experience designing and training large-scale deep learning models and working with large image datasets
• Strong mathematical foundations in machine learning and computer vision
• Ability to bridge research and production engineering
• Strong debugging and problem-solving skills

Nice to have

• Experience with perception for autonomous or embodied systems, such as SLAM, sensor fusion or state estimation

About the company

XYZ Reality created the world’s first engineering-grade augmented reality for construction. The company is a Series B business growing across the US, UK and Europe.