Juliana R.

Key Skills

Software

Appen

CloudFactory

CrowdSource

Data Annotation Tech

Dataloop

Datumbox

Mighty AI

OneForma

Scale AI

Toloka

Telus

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Text

Top Task Types

Routing

Segmentation

Text Generation

Text Summarization

Translation/Localization

Freelancer Overview

My experience is centered on being a dedicated and strategic AI professional, with a robust focus on LLM Safety and Alignment. I specialize in data curation and labeling, which is fundamental to creating high-quality training datasets. My core qualification is the proficiency in creating and organizing diverse, high-quality data in formats such as JSON and XLM. This data is essential for both fine-tuning and executing Red Teaming strategies. As an AI Safety Specialist, I developed and executed advanced Adversarial Prompting strategies , utilizing this data to simulate edge cases and ethical dilemmas, thereby proactively identifying vulnerabilities in LLMs. My expertise extends beyond text, encompassing the annotation and review of multimodal content (text, images, and videos). I have ensured the consistency and high quality of datasets , a crucial step for training and evaluating AI systems. This specialization in delivering high-granularity instruction and alignment training data, along with my ability to act as a Red Teamer mitigating high-level risks through safety assessments and content review, are the qualifications that distinguish me in the field of AI Alignment and Content Governance.

Labeling Experience

Red Teaming and Strategic Data Curation for LLM Alignment

Scale AITextRLHFRed Teaming

I served as a leader and specialist in the design and execution of advanced Adversarial Prompting (Red Teaming) strategies. This involved the creation and strategic curation of high-quality datasets in formats such as JSON and XLM, designed to simulate edge cases, ethical dilemmas, and high-risk scenarios. My focus was mitigating vulnerabilities and ensuring the ethical and technical compliance and robustness of the model. Additionally, I annotated and reviewed multimodal content (text, images, and videos), ensuring dataset consistency and adherence to platform policies and guidelines. This role included the translation and localization of technical materials (EN/ES/PT-BR) and the creation of detailed technical documentation.

I served as a leader and specialist in the design and execution of advanced Adversarial Prompting (Red Teaming) strategies. This involved the creation and strategic curation of high-quality datasets in formats such as JSON and XLM, designed to simulate edge cases, ethical dilemmas, and high-risk scenarios. My focus was mitigating vulnerabilities and ensuring the ethical and technical compliance and robustness of the model. Additionally, I annotated and reviewed multimodal content (text, images, and videos), ensuring dataset consistency and adherence to platform policies and guidelines. This role included the translation and localization of technical materials (EN/ES/PT-BR) and the creation of detailed technical documentation.

2024 - 2025

Education

Work History