Data Labelers- AV/ADAS

Role Overview

This is a hands-on opportunity for experienced data annotators to deepen their expertise in Autonomous Vehicle (AV) and Advanced Driver Assistance Systems (ADAS) labeling while preparing for production work with Digital Divide Data (DDD). Over an intensive two-week boot camp, you will refine the precision, reasoning, and quality-control skills that autonomous driving datasets demand — from point-cloud accuracy to temporal tracking. Trainees who meet the programme's rigorous standards may transition into live annotation projects as client work becomes available.

Key Responsibilities

• Annotate LiDAR scans and point-cloud data, including 2D and 3D bounding boxes, with high spatial accuracy.
• Perform image and video labeling tasks covering object detection, classification, and multi-frame tracking.
• Apply polygon, semantic, and instance segmentation techniques to define object boundaries and regions of interest.
• Label lane markings, road geometry, pedestrian activity, vehicle behaviour, and environmental features relevant to driving scenes.
• Interpret and consistently apply detailed annotation guidelines, adapting as specifications evolve.
• Track objects and agents across time, reasoning about ego-vehicle behaviour, traffic context, and interactions between road users.
• Carry out quality assurance checks on your own work and flag inconsistencies before submission.
• Complete language-based annotation tasks such as transcription, captioning, and prompt-response writing when required.

Requirements & Qualifications

• At least one year of hands-on experience in AV, ADAS, or closely related data annotation work.
• Practical familiarity with at least one of the following: LiDAR/point-cloud annotation, 2D/3D bounding boxes, image and video annotation, object detection/classification/tracking, polygon/semantic/instance segmentation, or AV/ADAS quality assurance.
• Strong grasp of annotation quality standards and the ability to follow detailed guidelines precisely.
• Excellent attention to detail and the ability to stay accurate across repetitive, complex tasks.
• Capacity to meet defined productivity and quality targets.
• Reading and writing proficiency in English.
• Not currently enrolled as a student.
• Willingness to complete experience verification and a practical skills assessment.
• Availability for potential project deployment after successfully completing the assessment process.

What the Training Covers

• AV fundamentals: scene understanding, object classification, LiDAR and point clouds, 3D annotation, precision work, and 2D/3D correlation.
• Driving and temporal reasoning: tracking over time, ego behaviour, traffic and road context, agent interaction, and spatio-temporal reasoning.
• Advanced reasoning: logical linking, causal and VLA reasoning, and working with evidence and uncertainty.
• Quality and adaptability: attention to detail, learning agility, adapting to changing guidelines, and independent QA.
• Hands-on application: practical annotation tasks, calibration, feedback, and exercises rather than classroom instruction alone.
• Post-training assessment: five evaluation tasks combining tool-based scoring and expert review, with critical gates for technical execution, reasoning, quality, and adaptability.
• Production readiness: trainees who meet the required standard move into production as project opportunities arise.

What We Offer / Why Join

The programme runs as a two-week, 80-hour in-person boot camp at the DDD Nairobi offices, blending theory, practical exercises, and daily skill checks. A training stipend is reimbursed to all successful trainees upon completion of the two-week boot camp. Trainees must complete the full programme and meet defined quality and proficiency standards to successfully finish it; successful completion does not guarantee immediate employment. Qualified participants join DDD's pre-screened talent bench and may be considered for future project assignments based on client demand and individual availability.