LiDAR 3D Annotation & Data Labeling Specialist
New
J
JobgetherArtificial Intelligence
BrazilContractMiddle
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Experience
- At least 6 months of hands-on experience
Requirements
- At least 6 months of hands-on experience with 3D LiDAR point cloud annotation, 3D semantic segmentation, or multi-sensor data labeling.
- Demonstrated proficiency with 3D annotation and spatial-data platforms such as Segments.ai, BasicAI, Cognic, Scale AI, CVAT, or equivalent tools.
- Strong understanding of 3D spatial geometry and the ability to interpret point-cloud data objectively.
- Proven ability to maintain high annotation accuracy, with experience working toward or achieving 95%+ quality standards preferred.
- Familiarity with 3D bounding boxes, object tracking, occlusion handling, spatial segmentation, and multi-frame consistency.
- Ability to accurately interpret object orientation and maintain consistency across pitch, roll, yaw, and heading vectors.
- Strong attention to detail and the ability to identify subtle spatial inconsistencies in dense datasets.
- Comfortable working with structured guidelines, quality benchmarks, and productivity expectations.
- Ability to use software shortcuts and hotkeys efficiently within 3D annotation environments.
- Stable internet connection and access to a capable PC or laptop suitable for 3D rendering and annotation workloads.
- Compatibility with required screen-recording software and project monitoring tools.
Responsibilities
- Create accurate 3D bounding boxes around vehicles, pedestrians, cyclists, static structures, and other relevant objects across LiDAR frame sequences.
- Perform detailed 3D semantic segmentation by labeling individual points within dense point clouds while avoiding gaps, overlaps, and inaccurate classifications.
- Review and refine AI-generated 3D annotations to ensure they meet established spatial accuracy and quality standards.
- Conduct multi-sensor quality assurance by validating LiDAR annotations against 2D camera feeds and checking sensor-fusion alignment.
- Track dynamic objects consistently across multiple LiDAR frames, maintaining accurate pitch, roll, yaw, and heading information.
- Apply strict cuboid boundary, point-density, occlusion, and spatial annotation guidelines throughout production workflows.
- Maintain a minimum target accuracy of 95% while balancing precision with project productivity requirements.
View Full Description & ApplyYou'll be redirected to the employer's site
