Solomon G.
Data Annotation Specialist — LiDAR & Semantic Segmentation (Autonomous Vehicle Perception)
Experience
Data Annotation Specialist — LiDAR & Semantic Segmentation (Autonomous Vehicle Perception)
Remotasks / Scale AI
- Performed frame-by-frame 3D segmentation and cuboid bounding on point clouds from multi-sensor AV systems (LiDAR + camera fusion).
- Classified and segmented dynamic and static objects (vehicles, pedestrians, cyclists, road features, infrastructure elements) with high precision and adherence to quality metrics.
- Consistently met stringent accuracy thresholds (QA audits and weekly performance reviews).
- Collaborated within distributed remote teams using annotation tools and guideline playbooks (Bee-LSS / Bee-LiDAR workflow).
- Contributed annotated datasets used for machine-learning model training by major tech clients in autonomous driving through Scale AI’s enterprise contracts.
- Analyzed and improved training datasets, enhancing model performance by 30% through meticulous data curation.
- Fostered a collaborative environment by working closely with developers and data scientists to ensure alignment on project goals.
- Cultivated a culture of innovation and collaboration, resulting in a dynamic work environment that encouraged creativity.
- Maintained accuracy scores above project targets using strict QA and revision workflows.
- Processed high volumes of consecutive frames while maintaining label consistency across sequences.
- Collaborated with global remote teams, tool specialists, and project auditors.
- Segmented objects in 3D environments including vehicles, pedestrians, cyclists, infrastructure, and road elements.
- Annotated and quality-checked over 2,500 LiDAR frames with accuracy above 96%.
- Ranked among top performers for segmentation accuracy and guideline compliance.
- Helped reduce rework rates through consistent annotation quality.
- Executed segmentation and classification of 3D point-cloud data for training AV perception algorithms.
- Applied object tracking logic across multiple temporal frames for continuity.
- Ensured dataset quality by identifying occlusions, no-label regions, and minimum LiDAR point rules.
Summary
Data Annotation Specialist with hands-on experience supporting autonomous-vehicle perception projects through LiDAR and semantic segmentation labeling.
Contributed high-accuracy datasets used by enterprise technology clients to improve ego-vehicle decision systems.
Dynamic and results-driven data annotator with over five years of experience in the AI and tech industry.
Proven track record of leveraging self-learning initiatives to enhance skills in data annotation and machine learning.
Passionate about startup culture, with a strong entrepreneurial spirit demonstrated through the creation and scaling of Dusty Games.
Committed to delivering high-quality, accurate datasets that contribute to project success and client satisfaction.
Languages
Education
Jomo Kenyatta University of Agriculture and Technology
Bachelor of Science · Business Innovation and Technology Management · Juja, Kenya
Similar Freelancers
Discover other experts with similar qualifications and experience