Bethany Frank

Key Skills

Software

Appen

CVAT

Data Annotation Tech

Deep Systems

Labelbox

LabelImg

Label Studio

SuperAnnotate

Telus

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Video

Top Task Types

Bounding Box

Classification

Evaluation/Rating

Prompt + Response Writing (SFT)

Segmentation

Freelancer Overview

With over three years of hands-on experience in data labeling, multimodal annotation, and AI training data development, I have contributed to high-quality datasets that support machine learning models across computer vision, NLP, and audio-visual domains. My work includes performing precise annotations such as bounding boxes, polygons, segmentation, activity tagging, NER, sentiment labeling, text classification, and structured rubric-based evaluations. I have collaborated closely with ML teams to interpret model weaknesses, refine guidelines, and enhance data quality through calibration reviews and workflow improvements. Throughout my career, I have demonstrated strong analytical judgment, consistency, and an ability to work with complex instructions in fast-paced, accuracy-driven environments. I specialize in multimodal tasks including Image-Text Matching, VQA, captioning, video tagging, reasoning verification, and dataset quality control. My background in QA for AI systems, together with experience reviewing multi-annotator submissions and producing gold-standard outputs, sets me apart as a detail-oriented expert capable of strengthening model performance through high-integrity data labeling.

Labeling Experience

Autonomous Vehicle Object Detection & Scene Segmentation Project

CVATVideoBounding BoxPolygon

I worked on a large-scale autonomous driving dataset focused on training computer vision models to detect vehicles, pedestrians, road objects, and environmental elements. My tasks included drawing bounding boxes and polygons for precise object localization, creating segmentation masks for lanes, sidewalks, and road surfaces, and tagging pedestrian and vehicle actions (e.g., turning, braking, crossing). I annotated over 22,000+ images and 3,500+ short video clips, maintaining quality scores above 96% across multiple QA reviews. I followed strict annotation guidelines, used calibrated label sets, and performed self-QC to ensure consistency across frames and scenes. The labeled dataset supported model improvement for hazard detection, path planning, and environmental understanding.

I worked on a large-scale autonomous driving dataset focused on training computer vision models to detect vehicles, pedestrians, road objects, and environmental elements. My tasks included drawing bounding boxes and polygons for precise object localization, creating segmentation masks for lanes, sidewalks, and road surfaces, and tagging pedestrian and vehicle actions (e.g., turning, braking, crossing). I annotated over 22,000+ images and 3,500+ short video clips, maintaining quality scores above 96% across multiple QA reviews. I followed strict annotation guidelines, used calibrated label sets, and performed self-QC to ensure consistency across frames and scenes. The labeled dataset supported model improvement for hazard detection, path planning, and environmental understanding.

2022 - 2022

Education

Work History