3D Object Detection Labeler

AI Trainer Jobs; United States; Remote; USD 76,000 - 138,000; Full time

Job summary

AI Trainer Jobs is seeking skilled annotators to label LiDAR data for autonomous vehicle projects. You will place precise 3D cuboids, classify objects, and estimate orientations in scenes captured from LiDAR sensors using browser-based tools such as Segments.ai or Scale LiDAR.

Precision is critical; mislabels can degrade downstream models. This role offers Expert tier compensation and potential senior labeler status with project lead responsibilities for high QA performers.

Qualifications

• LiDAR point cloud interpretation required.
• 3D cuboid annotation and object labeling per specs.
• Experience with CAD, GIS, or autonomous vehicle datasets preferred.

Responsibilities

• Annotate scenes with tight cuboids around objects.
• Classify objects and estimate orientation/heading.
• Follow per-project annotation specs and QA rules; respond within 24 hours.

Skills

LiDAR interpretation

3D cuboid annotation

Spatial reasoning

Object classification

Occlusion handling

Annotation tools

Tools

Segments.ai

Scale LiDAR

Job description

Pay: $55-100/hour

You'll annotate 3D point cloud data captured by LiDAR sensors, placing tight cuboid bounding boxes around objects like vehicles, pedestrians, cyclists, and static obstacles. Data originates from autonomous vehicle test drives and robotics platforms. Each scene requires correctly classifying object type, estimating orientation/heading, and flagging occluded or truncated objects. You'll work in tools like Scale AI's Lidar annotator, Segments.ai, or a custom internal web app — all browser-based with no local install needed.

Precision matters more than speed here. A misaligned cuboid or incorrect heading estimate on a vehicle directly degrades downstream perception model performance. You'll follow per-project annotation specs that define minimum point density thresholds for labeling, handling rules for partially visible objects, and class hierarchies. Disagreements with QA reviewers are resolved through an async comment thread, and you're expected to respond and correct within 24 hours.

This role suits people with spatial reasoning skills and experience in CAD, 3D modeling, GIS, or prior autonomous vehicle dataset work. Compensation is at the Expert tier given the technical precision involved. High-volume contributors with strong QA scores are eligible for senior labeler status, which includes project lead responsibilities and a pay bump.

Skills & requirements

• LiDAR point cloud interpretation
• 3D cuboid annotation
• Spatial reasoning
• Object classification
• Occlusion handling
• Annotation tool proficiency (Segments.ai / Scale LiDAR)

Category: Vision Annotation