Key Skills

Software

Internal/Proprietary Tooling

iMerit

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Bounding Box

Classification

Data Collection

Text Generation

Translation Localization

Freelancer Overview

I work at the intersection of psychology, language, and artificial intelligence. My background is international and interdisciplinary, and I’m most interested in how people think, communicate, and make sense of the world, especially across cultures. I bring that perspective to every project I join, whether I’m working with data, writing, or helping technology better understand human behavior. I have experience annotating social media content with psychological and cultural sensitivity to improve AI systems, particularly on Meta projects at iMerit. I speak Portuguese, English, and Spanish fluently, and I’ve lived and studied in Brazil, Mexico, and the United Kingdom, which helps me approach language and behavior with nuance and context. I enjoy making complex ideas clear and accessible, and I adapt easily to new tools, platforms, and team needs. My training in psychology and customer-focused communication helps me understand what really matters to users and how to make content and systems more effective.

Labeling Experience

Freelance Task Writer – Psychology Reasoning Evaluation

Internal Proprietary ToolingTextQuestion AnsweringText Generation

Sure! Here's your revised project description, tailored to directly answer the prompt in the image: --- This project involved the design and iterative refinement of psychology-based data labeling prompts aimed at testing reasoning vulnerabilities in large language models (LLMs). Tasks included crafting clear, specific, and cognitively challenging inputs to elicit targeted model failures, with each prompt structured to have a single correct response. The scope of the project integrated principles from developmental, cognitive, and social psychology to simulate realistic human interactions and evaluate model robustness. I contributed to large-scale alignment benchmarking and internal AI evaluation protocols. The project operated across multiple teams and followed strict quality standards, including internal review cycles and precision targets for prompt clarity, intent alignment, and reproducibility of model responses.

Sure! Here's your revised project description, tailored to directly answer the prompt in the image: --- This project involved the design and iterative refinement of psychology-based data labeling prompts aimed at testing reasoning vulnerabilities in large language models (LLMs). Tasks included crafting clear, specific, and cognitively challenging inputs to elicit targeted model failures, with each prompt structured to have a single correct response. The scope of the project integrated principles from developmental, cognitive, and social psychology to simulate realistic human interactions and evaluate model robustness. I contributed to large-scale alignment benchmarking and internal AI evaluation protocols. The project operated across multiple teams and followed strict quality standards, including internal review cycles and precision targets for prompt clarity, intent alignment, and reproducibility of model responses.

2023

Portuguese Language Specialist – Text annotation and safety

ImeritTextBounding BoxClassification

This project involved annotating and evaluating AI-generated text outputs in Portuguese across social media platforms, ensuring alignment with Meta’s content, safety, and cultural guidelines. Annotation tasks included labeling text-based content for hate speech, harassment, misinformation, and psychological risk factors, with attention to linguistic nuance and regional variations. A key part of the scope was identifying ethical and psychological risks in moderation workflows, particularly content potentially harmful to vulnerable populations, and contributing insights to improve AI detection systems. The project was large-scale and collaborative, involving multiple language specialists and product teams. Quality was ensured through structured review cycles, accuracy benchmarks, and alignment testing, with regular feedback loops to refine training inputs and support responsible policy implementation.

This project involved annotating and evaluating AI-generated text outputs in Portuguese across social media platforms, ensuring alignment with Meta’s content, safety, and cultural guidelines. Annotation tasks included labeling text-based content for hate speech, harassment, misinformation, and psychological risk factors, with attention to linguistic nuance and regional variations. A key part of the scope was identifying ethical and psychological risks in moderation workflows, particularly content potentially harmful to vulnerable populations, and contributing insights to improve AI detection systems. The project was large-scale and collaborative, involving multiple language specialists and product teams. Quality was ensured through structured review cycles, accuracy benchmarks, and alignment testing, with regular feedback loops to refine training inputs and support responsible policy implementation.

2022

Portuguese Language Specialist – Image annotation and safety

ImeritImageBounding BoxClassification

Certainly! Here is your revised project description tailored to the prompt in the image, with a focus on image-based tasks: --- This project focused on annotating and evaluating AI-generated outputs in Portuguese, specifically image and image-text content across major social media platforms. Tasks involved labeling visual data in alignment with Meta’s content moderation, safety, and cultural guidelines, ensuring compliance with regional norms and linguistic nuance. The scope included identifying ethical and psychological risks, particularly in content affecting vulnerable populations, such as children and individuals with mental health concerns. I contributed to refining detection systems by flagging subtle visual cues and context-dependent imagery that could lead to harm or misinformation. The project was part of a large-scale, multilingual operation, and adhered to strict quality measures, including daily accuracy thresholds, inter-annotator agreement checks, and weekly audits.

Certainly! Here is your revised project description tailored to the prompt in the image, with a focus on image-based tasks: --- This project focused on annotating and evaluating AI-generated outputs in Portuguese, specifically image and image-text content across major social media platforms. Tasks involved labeling visual data in alignment with Meta’s content moderation, safety, and cultural guidelines, ensuring compliance with regional norms and linguistic nuance. The scope included identifying ethical and psychological risks, particularly in content affecting vulnerable populations, such as children and individuals with mental health concerns. I contributed to refining detection systems by flagging subtle visual cues and context-dependent imagery that could lead to harm or misinformation. The project was part of a large-scale, multilingual operation, and adhered to strict quality measures, including daily accuracy thresholds, inter-annotator agreement checks, and weekly audits.

2022

Education

Work History