Dayana Duffus Miranda

Key Skills

Software

Remotasks

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Prompt Response Writing SFT

Question Answering

Text Generation

Translation Localization

Freelancer Overview

AI training specialist with a PhD in Economic Sciences and four years of experience assessing large-language-model outputs. On Outlier’s 2024-2025 project I reviewed 12 k+ Spanish- and English-language prompt-response pairs—ranking alternatives, tagging error types, and suggesting rewrites that raised overall model quality and user relevance. My background as a bilingual professor and copywriter lets me detect nuance, cultural bias, and style issues that automated checks overlook. While my focus has been expert judgment rather than tool operation, I adapt quickly to new annotation interfaces and apply rigorous QA standards to every dataset I evaluate.

Labeling Experience

LLM Prompt-Response Evaluation – Outlier (2024-2025)

Scale AITextQuestion Answering

Contract evaluator reviewing 12 k+ bilingual (EN/ES) prompt-response pairs. Tasks: ranked alternatives, tagged error types (factuality, style, bias), and drafted improved answers. Achieved ≥ 98 % agreement with gold labels and helped raise overall model acceptance rate by 15 %.

Contract evaluator reviewing 12 k+ bilingual (EN/ES) prompt-response pairs. Tasks: ranked alternatives, tagged error types (factuality, style, bias), and drafted improved answers. Achieved ≥ 98 % agreement with gold labels and helped raise overall model acceptance rate by 15 %.

2024 - 2024

Education

Work History