Tara K.

Key Skills

Software

Don't disclose

Top Subject Matter

AI model evaluation and dataset labeling for LLM training

Content moderation and classification for social media data

Legal Services & Contract Review

Top Data Types

Text

Audio

Image

Document

Top Task Types

Classification

Freelancer Overview

Freelance AI Trainer & Data Quality Specialist | Prolific & CloudResearch Connect. Brings 21+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Don't disclose. Education includes Coursework, Des Moines Area Community College and Coursework, University of Arizona Global Campus. AI-training focus includes data types such as Text and Audio and labeling workflows including Evaluation, Rating, and Classification.

Labeling Experience

Freelance AI Trainer & Data Quality Specialist | Prolific & CloudResearch Connect

Don't discloseText

Evaluated AI-generated responses for factual accuracy, semantic reasoning, safety, constraints, and overall helpfulness while ensuring strict data integrity under evolving guidelines. Compared and ranked multiple LLM outputs using rigorous evaluation matrices to support model training data improvements. Performed dataset annotation, categorization, and quality assurance tasks aligned to project requirements. • AI model evaluation across accuracy, reasoning, safety, and helpfulness • Dataset labeling and quality assurance • LLM output comparison and ranking using evaluation matrices • Prompt authorship and refinement to stress-test model behavior

2024 - Present

Social Media Content Moderator - Respondology

AudioClassification

Monitored social media comment streams to protect brand reputation for high-profile corporate clients, athletes, and influencers. Identified and filtered toxic content, hate speech, and spam while preserving genuine user engagement. Applied rapid judgment using evolving brand guidelines and contextual signals to classify content accurately and consistently. • Actively moderated high-volume comment and engagement data streams • Detected and mitigated emerging automated bot networks and spam patterns • Followed complex brand and community standards for content classification • Adjusted moderation decisions based on nuance and guideline updates without suppressing legitimate users

2021 - Present

Social Media Moderator | Respondology

Don't discloseTextClassification

Moderated social media comment streams by filtering toxic content, hate speech, and spam for high-profile corporate clients and influencers. Reviewed high-volume data streams to identify and mitigate automated bot networks and emerging spam patterns. Classified content according to shifting brand guidelines and contextual nuances while minimizing impact on genuine user engagement. • Toxic/hate/spam detection in social media text • Bot and spam pattern identification • Content classification under evolving guidelines • Quality-focused moderation to protect brand reputation

2021 - Present

Education

Work History