Diana Zagrean

Key Skills

Software

Clickworker

CVAT

Doccano

Labelbox

LightTag

OneForma

Prodigy

Remotasks

Scale AI

SuperAnnotate

Other

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Document

Image

Text

Top Task Types

Classification

Emotion Recognition

Fine Tuning

RLHF

Translation Localization

Freelancer Overview

I am an experienced AI training data contributor with a strong background in text evaluation, prompt ranking, translation quality assessment, and relevance judgment. I have worked on projects with platforms like Outlier.ai, UHRS (via Clickworker), and Appen, where I focused on evaluating AI-generated content, ranking chatbot outputs, and analyzing tone, sentiment, and factual consistency across multiple languages. My work has included detailed rationales, rule-based annotation, and structured feedback in both short-form and long-form tasks. Fluent in English, German, Romanian, Spanish, and French, I bring a multilingual and cross-cultural perspective to evaluation tasks. I also have a background in journalism and psychology, which supports my ability to assess nuance, bias, clarity, and user intent. I’m comfortable adapting to detailed guidelines and enjoy working on tasks that require both linguistic sensitivity and logical precision—especially those related to LLM alignment, content moderation, or multilingual evaluation.

Labeling Experience

Freelance Visual Data Annotator – Search Relevance & Image Classification

AppenImageEntity Ner ClassificationClassification

Contributed to image annotation and visual relevance projects via UHRS and OneForma. Tasks included labeling product images, evaluating ad alignment with user queries, classifying inappropriate or low-quality visuals, and selecting the most relevant images in side-by-side comparisons. Work required attention to visual context, consistency, and precise judgment under shifting guidelines. Also handled basic content moderation to flag harmful or misleading images.

Contributed to image annotation and visual relevance projects via UHRS and OneForma. Tasks included labeling product images, evaluating ad alignment with user queries, classifying inappropriate or low-quality visuals, and selecting the most relevant images in side-by-side comparisons. Work required attention to visual context, consistency, and precise judgment under shifting guidelines. Also handled basic content moderation to flag harmful or misleading images.

2023

Freelance AI Evaluator – Multilingual Text & Prompt Rating

OtherTextClassificationEmotion Recognition

Worked on AI training and evaluation projects involving large language model outputs, including ranking chatbot responses (RLHF), evaluating multilingual prompt translations, tone and sentiment classification, and labeling factual consistency. Tasks required strict adherence to guidelines, detailed rationales, and linguistic accuracy. Contributed to ongoing model refinement by comparing AI-generated answers and identifying preferred completions in English, German, Romanian, and Spanish.

Worked on AI training and evaluation projects involving large language model outputs, including ranking chatbot responses (RLHF), evaluating multilingual prompt translations, tone and sentiment classification, and labeling factual consistency. Tasks required strict adherence to guidelines, detailed rationales, and linguistic accuracy. Contributed to ongoing model refinement by comparing AI-generated answers and identifying preferred completions in English, German, Romanian, and Spanish.

2023

Education

Work History