Breanna Downes

Key Skills

Software

AWS SageMaker

Appen

Argilla

Clickworker

CrowdSource

CVAT

Data Annotation Tech

Dataloop

LabelImg

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

3D Sensor

Computer Code Programming

Image

Top Task Types

Bounding Box

Classification

Computer Programming Coding

Data Collection

Text Generation

Freelancer Overview

Experienced AI Trainer specializing in LLM evaluation, text generation assessment, and computer vision annotation. I have worked extensively on data labeling, AI model fine-tuning, and quality assurance across multiple domains, including natural language processing (NLP), autonomous vehicle datasets, and multilingual sentiment analysis. With hands-on expertise in tools like CVAT, Labelbox, and Remotasks, I have contributed to the development of high-quality training data for machine learning models. My ability to ensure data accuracy, detect biases, and provide insightful feedback helps improve AI performance. Passionate about advancing AI through precise annotations and in-depth evaluations, I excel in multilingual AI projects and complex data labeling tasks.

Labeling Experience

LLM Evaluation and AI Model Fine-Tuning

Label StudioTextClassificationText Generation

As part of an AI model fine-tuning project, I evaluated responses generated by large language models (LLMs) to ensure accuracy, coherence, and alignment with human intent. The work involved: Ranking model outputs based on factual correctness, fluency, and relevance. Classifying responses for bias detection, sentiment analysis, and content safety. Writing and refining prompt-response pairs to enhance model learning. Contributing to reinforcement learning (RLHF) by training models with human feedback loops.

As part of an AI model fine-tuning project, I evaluated responses generated by large language models (LLMs) to ensure accuracy, coherence, and alignment with human intent. The work involved: Ranking model outputs based on factual correctness, fluency, and relevance. Classifying responses for bias detection, sentiment analysis, and content safety. Writing and refining prompt-response pairs to enhance model learning. Contributing to reinforcement learning (RLHF) by training models with human feedback loops.

2023

Autonomous Vehicle Data Annotation

CVATImageBounding BoxPolygon

Worked on a self-driving car dataset involving large-scale image and video annotation for object detection, pedestrian tracking, and lane recognition. Tasks included: Annotating vehicles, traffic signs, road markings, and pedestrians using bounding boxes and segmentation. Labeling multi-frame video sequences for motion tracking and action recognition. Ensuring data consistency and accuracy through quality control checks. The dataset contributed to training AI-powered navigation and perception systems for autonomous vehicles.

Worked on a self-driving car dataset involving large-scale image and video annotation for object detection, pedestrian tracking, and lane recognition. Tasks included: Annotating vehicles, traffic signs, road markings, and pedestrians using bounding boxes and segmentation. Labeling multi-frame video sequences for motion tracking and action recognition. Ensuring data consistency and accuracy through quality control checks. The dataset contributed to training AI-powered navigation and perception systems for autonomous vehicles.

2022 - 2023

Education

Work History