Micaela Entivero

Key Skills

Software

Mercor

Labelbox

Top Subject Matter

AI-generated content evaluation

adversarial prompt analysis

translation validation

Top Data Types

Text

Image

Document

Top Task Types

Bounding Box

Text Summarization

Transcription

RLHF

Evaluation Rating

Question Answering

Text Generation

Entity Ner Classification

Classification

Object Detection

Freelancer Overview

I am an AI Evaluation Specialist with experience reviewing and annotating training data for large language models across text and multimodal formats. In my current role, I analyze AI-generated outputs to identify logical inconsistencies, translation inaccuracies, and structural issues, ensuring high-quality datasets for model improvement. My work involves applying detailed guidelines, maintaining consistency across large volumes of data, and documenting findings through structured annotations. I bring strong attention to detail, analytical thinking, and experience working with bilingual (English–Spanish) data. While my background is primarily focused on text and multimodal evaluation, I am highly adaptable and familiar with annotation workflows, quality control processes, and data labeling standards. I am comfortable learning new tools quickly and applying precise criteria, which allows me to transition effectively into image annotation and bounding box tasks.

Labeling Experience

AI Generalist (Spanish & Logic Evaluation)

MercorText

Reviewed AI-generated prompts and responses across multilingual datasets to identify logic flaws, translation issues, and bias risks. Applied filtering guidelines to detect and redact personal identifiable information (PII) and ensure safety compliance. Produced structured annotations and reports to document failure patterns and refine training data quality. • Conducted bilingual (EN–ES) validation for semantic and contextual accuracy. • Evaluated multiple output formats including text, image, and video. • Assessed prompt and response quality within standardized frameworks. • Enhanced AI datasets by prompt refinement and annotation.

Reviewed AI-generated prompts and responses across multilingual datasets to identify logic flaws, translation issues, and bias risks. Applied filtering guidelines to detect and redact personal identifiable information (PII) and ensure safety compliance. Produced structured annotations and reports to document failure patterns and refine training data quality. • Conducted bilingual (EN–ES) validation for semantic and contextual accuracy. • Evaluated multiple output formats including text, image, and video. • Assessed prompt and response quality within standardized frameworks. • Enhanced AI datasets by prompt refinement and annotation.

2025 - Present

AI Generalist (Spanish & Logic Evaluation)

Text

Evaluated AI-generated outputs in text, audio, image, and video formats to detect logical inconsistencies, translation inaccuracies, and bias risks. Performed adversarial prompt analysis and bilingual (EN–ES) validation to ensure semantic accuracy and contextual alignment. Produced structured annotations regarding reasoning flaws, jailbreak behaviors, and risk areas to improve AI training datasets and model quality. • Applied structured frameworks and documented criteria for consistent evaluations. • Reviewed text, audio, image, and video content for weakness patterns. • Identified translation and logic errors across multimodal outputs. • Created annotations supporting dataset refinement and AI model improvement.

Evaluated AI-generated outputs in text, audio, image, and video formats to detect logical inconsistencies, translation inaccuracies, and bias risks. Performed adversarial prompt analysis and bilingual (EN–ES) validation to ensure semantic accuracy and contextual alignment. Produced structured annotations regarding reasoning flaws, jailbreak behaviors, and risk areas to improve AI training datasets and model quality. • Applied structured frameworks and documented criteria for consistent evaluations. • Reviewed text, audio, image, and video content for weakness patterns. • Identified translation and logic errors across multimodal outputs. • Created annotations supporting dataset refinement and AI model improvement.

2025 - Present

Education

Work History