LiDAR / 3D Point Cloud Perception Engineer

We are developing SmartView, a vehicle-based geospatial monitoring platform that combines LiDAR, RGB cameras, GNSS and IMU data to detect, measure and geographically position objects and conditions in public-space environments. We already have specialists covering computer vision, localization/GNSS, GIS and backend processing. We are specifically looking for someone to own the LiDAR / 3D point-cloud perception layer. This is not primarily a SLAM or GIS role. We need someone who can turn raw LiDAR point clouds into structured object-level measurements and positions. Scope of Work You would work on tasks such as: Processing raw LiDAR point clouds from a moving vehicle. Filtering and preprocessing point-cloud data. Ground/non-ground separation. Point-cloud clustering and feature extraction. Detecting or segmenting objects/features in 3D. Calculating object-level properties such as: XYZ centroid distance/range height width/dimensions 3D geometry/bounding regions Developing geometric LiDAR detectors where appropriate. Producing consistent object positions in the sensor/vehicle coordinate frame. Supporting camera ↔ LiDAR association where required. Working with our localization engineer to transform these observations into world coordinates. Building repeatable Python/C++ point-cloud processing pipelines. Our initial use cases include vegetation, roadside objects, obstacles and other public-space features. Strong Fit If You Have Experience With PCL, Open3D, PDAL or equivalent point-cloud libraries Python and/or C++ LiDAR point-cloud processing 3D feature/object extraction clustering and segmentation mobile mapping geometric measurements from point clouds LiDAR coordinate frames and sensor extrinsics Experience with Hesai, Ouster, Velodyne or similar 3D LiDAR sensors is particularly relevant. Camera-LiDAR fusion, ROS2 and GNSS/IMU experience are useful, but not mandatory. Those areas are already supported by other specialists in the team. Initial Engagement The first objective will be to demonstrate a reliable chain such as: LiDAR point cloud → selected object/feature → 3D geometry → XYZ position in vehicle frame → structured output If successful, the role can expand into the wider SmartView processing pipeline. When Applying Please answer these questions specifically. Generic responses will not be considered: 1. Describe one project where you personally processed raw 3D LiDAR point clouds. Which sensor and libraries did you use? 2. Have you implemented object clustering, 3D object detection, feature extraction or point-cloud segmentation? Give one concrete example. 3. Have you automatically calculated object properties such as centroid, XYZ position, height, dimensions or 3D bounding boxes? 4. Have you worked with mobile-mapping LiDAR or point clouds combined with GNSS/IMU trajectories? 5. Have you worked with camera + LiDAR data? If yes, briefly explain what you implemented. 6. Please provide a relevant GitHub repository, video, screenshot, paper or project example if available. Important: We are looking for hands-on implementation experience, not only experience viewing, annotating or manually classifying point clouds.