James Santillana

Key Skills

Software

Scale AI

Mercor

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Video

Top Task Types

Evaluation Rating

Fine Tuning

Prompt Response Writing SFT

RLHF

Text Generation

Freelancer Overview

I have hands-on experience in data labeling and AI training data processes, having worked remotely as a Generalist and Reviewer for leading AI companies. My work involved daily annotation and systematic evaluation of multimodal AI outputs—including images, audio, and video—using established labeling platforms and detailed guidelines to ensure high-quality, precise data. I have trained and evaluated large language models (LLMs), utilizing prompt engineering, annotation tools, and rigorous rubrics to assess and improve model responses. My background also includes reviewing and providing feedback on peer submissions to maintain annotation quality. With strong technical fundamentals, attention to detail, and fluency in both English and Spanish, I am adept at supporting projects in computer vision, NLP, and more.

Labeling Experience

Generalist

MercorImageRLHFEvaluation Rating

• Performed daily preference ranking of hundreds of multimodal AI outputs (images, audio, and video) against prompts to evaluate model performance and alignment. • Applied systematic comparative judgment techniques to assess relevance, quality, and coherence of generative model outputs. • Utilized labeling platforms and guidelines to ensure precise annotation.

• Performed daily preference ranking of hundreds of multimodal AI outputs (images, audio, and video) against prompts to evaluate model performance and alignment. • Applied systematic comparative judgment techniques to assess relevance, quality, and coherence of generative model outputs. • Utilized labeling platforms and guidelines to ensure precise annotation.

2025 - 2025

Generalist

Scale AITextText GenerationFine Tuning

• Trained LLMs (Large Language Models) utilizing annotation tools, and prompt engineering interfaces. • Evaluated, and graded model responses based on provided rubric. • Reviewed submissions by other attempters, provided feedback, and revised inadequate input.

• Trained LLMs (Large Language Models) utilizing annotation tools, and prompt engineering interfaces. • Evaluated, and graded model responses based on provided rubric. • Reviewed submissions by other attempters, provided feedback, and revised inadequate input.

2024 - 2025

Education

Work History