Irna J.

Key Skills

Software

Other

Top Subject Matter

Government policy and public communications in Indonesian/English

Public relations content quality

misinformation detection

Top Data Types

Text

Video

Top Task Types

RLHF

Freelancer Overview

Social Media Manager & Government Public Relations Officer — content screening and evaluation for AI-style output qualit. Brings 7+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes English Education Degree, State University of Medan (2014). AI-training focus includes data types such as Text and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

Content evaluation and misinformation detection embedded in government PR duties (transferable to AI output rating)

OtherText

Performed systematic content evaluation of official posts and communications by checking language precision, framing, and compliance with government standards. Assessed audience sentiment and the potential impact of messages to support better quality judgments. Applied prompt understanding concepts by verifying that communication outputs match intended objectives and user/public needs. • Screened content for factual accuracy and clarity • Evaluated tone and potential misinterpretation • Ensured consistency across cross-department messaging • Supported crisis situations with rapid reliability judgments

2020 - Present

Social Media Manager & Government Public Relations Officer — content screening and evaluation for AI-style output quality assessment (AI Trainer background)

OtherTextRLHF

Provided AI training and evaluation support by assessing written outputs for accuracy, clarity, tone, and helpfulness aligned with RLHF workflows. Identified misinformation, bias, and misleading framing in public communications to improve the reliability of model responses. Reviewed and refined candidate text outputs and prompts to ensure they meet intended user intent and domain requirements. • Evaluated responses for factual accuracy and potential public misinterpretation • Flagged politically sensitive or harmful content for safer generation • Composed structured written feedback to guide response improvement • Monitored public discourse to detect misinformation narratives for model QA

2020 - Present

Education

Work History