Perception Engineer

Requirements

• Experience with SLAM, occupancy mapping, or related localisation techniques
• Familiarity with PCL (Point Cloud Library) or Open3D
• Familiarity with foundation vision models for segmentation or detection (e.g. SAM2) and how to apply them in production contexts
• Exposure to MLOps tooling - experiment tracking (MLflow, W&B), dataset versioning, model deployment
• Academic or project work in robotics, autonomous systems, or applied ML
• Interest in warehouse logistics, supply chain, or industrial automation

Benefits

Additional Information

Perception Engineer / LiDAR / Python / C++ / Computer Vision / Location: Wallingford - Oxfordshire - Hybrid Permanent At Dexory, we are transforming the warehousing and logistics industry through data intelligence. We're currently recruiting for a number of perception engineers. We are looking for someone who is curious, technically sharp, and comfortable operating at the boundary between research and production. You do not need to have worked in robotics or logistics before. What matters is that you are strong on fundamentals - 3D geometry, signal processing, computer vision, or machine learning - and motivated by the challenge of making perception systems work reliably in messy, real-world environments. Offices based in Wallingford, Oxford but open to flexible set ups. Please apply for more information. Responsibilities / Skills - Perception pipeline development: develop and improve systems for point cloud processing, object detection, and scene understanding in warehouse environments, working across both classical and learning-based approaches. - Sensor data analysis: work directly with 3D LiDAR scans and camera streams to debug, evaluate, and improve system behaviour across a growing fleet of customer sites. - Data collection and labelling: contribute to workflows that build and maintain the datasets underpinning both classical and learned perception components. - Cross-team collaboration: work closely with robotics engineers to integrate perception capabilities into the broader ROS2-based autonomy stack and ensure robust end-to-end behaviour. - Experimentation and evaluation: run structured experiments, track results, and bring rigour to how we measure accuracy improvements and regressions across diverse warehouse environments. - Tooling and debugging: build and maintain tools to support analysis, visualisation, and debugging of perception system performance in the field. - Technical engagement: stay current with relevant research and emerging techniques, sharing findings with the team and contributing to the perception roadmap over time. We expect a solid foundation in one or more of the following areas: - 3D perception, point cloud processing, or sensor fusion (e.g. LiDAR, depth cameras, IMU) - Computer vision using classical or learning-based methods (OpenCV, PyTorch, TensorFlow) - Machine learning or deep learning, particularly applied to spatial or structured data - Robotics development, including familiarity with ROS or ROS2 Across all backgrounds, we expect: - Strong Python skills, with C++ a meaningful advantage given the performance-critical nature of the work. - Solid engineering fundamentals: clean, testable code, careful attention to edge cases, and genuine ownership of quality. - A practical mindset: comfortable bridging theory and implementation, and able to make progress without perfect data or perfect problem definitions. - Good communication: able to explain technical reasoning clearly to colleagues with different backgrounds and engage with feedback constructively You should be comfortable travelling to customer warehouse sites, domestically and internationally, to observe system behaviour and gather ground truth data in real deployment environments.

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at dexory? Share your experience