Carol Tortoza

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Prompt Response Writing SFT

Freelancer Overview

I bring over three years of experience in remote project coordination and virtual assistance, with a strong focus on organizing, managing, and processing data to support business operations. My background includes extensive hands-on work with spreadsheets, digital tools, and workflow automation platforms like ClickUp and Zapier, which I have used to streamline repetitive tasks and improve data accuracy and team collaboration. I am highly detail-oriented and comfortable working across time zones, consistently ensuring data is well-organized and easily accessible for teams. My ability to quickly adapt to new technologies and anticipate team needs allows me to efficiently handle data annotation, labeling, and quality control tasks, making me a reliable partner for AI training data projects in creative, e-commerce, and content-driven environments.

Labeling Experience

Data Annotation

LabelboxTextPrompt Response Writing SFT

Project Scope: The project focused on evaluating LLM-generated responses to improve model alignment and reduce hallucinations. The goal was to identify the most accurate and contextually appropriate response to a given prompt. Specific Data Labeling Task: For each task, I was given one prompt and three model-generated responses. I analyzed them and selected the most accurate, coherent, and factually correct answer, identifying hallucinations or logical inconsistencies in the other responses. Project Size: I completed multiple batches of prompt-response evaluations across several annotation sessions, handling a high volume of comparative assessments. Quality Measures Adhered To: All work followed strict annotation guidelines, prioritizing factual accuracy, logical consistency, clarity, and relevance. Consistency and attention to detail were essential to meet quality standards.

Project Scope: The project focused on evaluating LLM-generated responses to improve model alignment and reduce hallucinations. The goal was to identify the most accurate and contextually appropriate response to a given prompt. Specific Data Labeling Task: For each task, I was given one prompt and three model-generated responses. I analyzed them and selected the most accurate, coherent, and factually correct answer, identifying hallucinations or logical inconsistencies in the other responses. Project Size: I completed multiple batches of prompt-response evaluations across several annotation sessions, handling a high volume of comparative assessments. Quality Measures Adhered To: All work followed strict annotation guidelines, prioritizing factual accuracy, logical consistency, clarity, and relevance. Consistency and attention to detail were essential to meet quality standards.

2025

Education

Work History