David Resendez

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

3D Sensor

Audio

Image

Text

Top Label Types

Bounding Box

Segmentation

Cuboid

Freelancer Overview

I bring over 15 years of experience working at the intersection of cybersecurity, Trust & Safety, and AI content moderation for leading global technology platforms. My expertise includes hands-on review and enforcement of platform and AI policies, large-scale data analysis to identify abuse and misinformation, and direct collaboration with engineering teams to enhance AI-assisted moderation and human-in-the-loop workflows. I have a strong track record in data annotation and labeling for sensitive user-generated content, ensuring high-quality training data for responsible AI systems in domains such as social media, advertising, and cloud security. Proficient in Python and SQL, I leverage trend analysis, investigative reporting, and rigorous compliance standards (GDPR, ISO 27001, NIST) to support the development of ethical, reliable AI models and data pipelines. My approach centers on ethical judgment, policy-consistent enforcement, and continuous improvement in data quality and user safety.

Labeling Experience

Autonomous Driving 2D/3D Image & LiDAR Annotation for Object Detection

CVAT3D SensorBounding BoxSegmentation

Led high precision data annotation for an autonomous driving AI system involving over 250,000 images and 8,000+ LiDAR frames. The project included 2D bounding box annotation for vehicles, pedestrians, cyclists, and traffic signs, as well as 3D cuboid labeling using LiDAR point clouds. Performed semantic and instance segmentation for road elements, lane markings, and dynamic objects. Managed frame-by-frame video object tracking to improve temporal consistency in model training. Implemented multi stage quality assurance workflows, including gold-standard validation sets, inter-annotator agreement measurement, and edge case review sessions (occlusion, motion blur, weather conditions). Contributed to annotation guideline development and team calibration sessions to maintain labeling consistency across a 25+ annotator team. The refined dataset improved object detection model performance (mAP increased by 18%) and reduced label noise significantly, resulting in better detection accuracy.

Led high precision data annotation for an autonomous driving AI system involving over 250,000 images and 8,000+ LiDAR frames. The project included 2D bounding box annotation for vehicles, pedestrians, cyclists, and traffic signs, as well as 3D cuboid labeling using LiDAR point clouds. Performed semantic and instance segmentation for road elements, lane markings, and dynamic objects. Managed frame-by-frame video object tracking to improve temporal consistency in model training. Implemented multi stage quality assurance workflows, including gold-standard validation sets, inter-annotator agreement measurement, and edge case review sessions (occlusion, motion blur, weather conditions). Contributed to annotation guideline development and team calibration sessions to maintain labeling consistency across a 25+ annotator team. The refined dataset improved object detection model performance (mAP increased by 18%) and reduced label noise significantly, resulting in better detection accuracy.

2023 - 2024

Education

Work History