Rosemary Barrios

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Task Types

Bounding Box

Cuboid

Data Collection

Polygon

Segmentation

Freelancer Overview

Highly dedicated Data Annotation and Labeling Specialist with 3 years of experience executing critical projects for Artificial Intelligence (AI) and Machine Learning (ML) models. Proven expertise in handling 2D, 3D (LiDAR), and semantic data across over 20 projects within high-tech sectors, particularly Autonomous Vehicles and Geospatial Mapping. Committed to high-quality output, precision, and strict adherence to project guidelines.

Labeling Experience

Critical Road Feature Labeling for AI Training

Scale AIImageSegmentation

Executed high-precision Roadway Semantic Segmentation for HD mapping projects, focusing on accurate delineation of complex urban and suburban environments. Task required granular labeling of all road elements, including traffic lanes (type and direction), curbs, medians, sidewalks, pedestrian crossings, and road surface boundaries to train robust models for autonomous navigation and infrastructure monitoring.

Executed high-precision Roadway Semantic Segmentation for HD mapping projects, focusing on accurate delineation of complex urban and suburban environments. Task required granular labeling of all road elements, including traffic lanes (type and direction), curbs, medians, sidewalks, pedestrian crossings, and road surface boundaries to train robust models for autonomous navigation and infrastructure monitoring.

2024 - 2025

Multi-Frame 3D Cuboid Annotation and Attribute Classification for Automotive Datasets

Scale AIVideoCuboid

Executed specialized 3D Cuboid Annotation within a video context (VideoBox environment), performing accurate spatial and temporal bounding of vehicles across extended sequences (minimum 50 frames per instance). Tasks included precise object tracking and rigorous attribute classification, identifying criteria such as: Vehicle Type (e.g., sedan, truck, bus), Operational Status (lights on/off), and Activity Status (parked, stopped, in motion). This high-fidelity annotation directly contributed to training sophisticated models for real-time object recognition and behavioral prediction in autonomous driving systems.

Executed specialized 3D Cuboid Annotation within a video context (VideoBox environment), performing accurate spatial and temporal bounding of vehicles across extended sequences (minimum 50 frames per instance). Tasks included precise object tracking and rigorous attribute classification, identifying criteria such as: Vehicle Type (e.g., sedan, truck, bus), Operational Status (lights on/off), and Activity Status (parked, stopped, in motion). This high-fidelity annotation directly contributed to training sophisticated models for real-time object recognition and behavioral prediction in autonomous driving systems.

2024 - 2024

Dynamic Environmental Classification for Autonomous Driving Safety

Scale AIVideoSegmentation

Executed specialized Video Frame Annotation and Classification on sequences captured by vehicle-mounted cameras. The project involved processing continuous video streams (over 50 frames per sequence) to dynamically segment and classify environmental and visibility conditions crucial for autonomous driving safety models. Key tasks included detailed semantic labeling and categorization of variable factors such as: Weather Conditions: Rain, heavy mist/fog, clear skies, and snow conditions. Visual Impairments: Lens flares (direct sunlight), windshield obstructions (dirt, water droplets), and low-light conditions. Visibility Assessment: Categorizing the overall scene visibility on a frame-by-frame basis to train AI in making real-time hazard assessments.

Executed specialized Video Frame Annotation and Classification on sequences captured by vehicle-mounted cameras. The project involved processing continuous video streams (over 50 frames per sequence) to dynamically segment and classify environmental and visibility conditions crucial for autonomous driving safety models. Key tasks included detailed semantic labeling and categorization of variable factors such as: Weather Conditions: Rain, heavy mist/fog, clear skies, and snow conditions. Visual Impairments: Lens flares (direct sunlight), windshield obstructions (dirt, water droplets), and low-light conditions. Visibility Assessment: Categorizing the overall scene visibility on a frame-by-frame basis to train AI in making real-time hazard assessments.

2024 - 2024

3D Cuboid Annotation for Static Traffic Sign Recognition

Scale AIVideoCuboid

Executed specialized 3D Cuboid Annotation on static road assets, focusing exclusively on traffic signage within dynamic video sequences (approximately 20 frames per sequence). The task involved creating precise 3D boundary boxes around various types of regulatory, warning, and informational signs to ensure accurate identification and localization in training data. This work directly supported the development of robust visual recognition systems for road infrastructure used in autonomous navigation and HD mapping initiatives.

Executed specialized 3D Cuboid Annotation on static road assets, focusing exclusively on traffic signage within dynamic video sequences (approximately 20 frames per sequence). The task involved creating precise 3D boundary boxes around various types of regulatory, warning, and informational signs to ensure accurate identification and localization in training data. This work directly supported the development of robust visual recognition systems for road infrastructure used in autonomous navigation and HD mapping initiatives.

2023 - 2023

Education

Work History