Chelsie Turner

Key Skills

Software

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Video

Text

Audio

Top Label Types

Question Answering

Object Detection

Action Recognition

Tracking

RLHF

Red Teaming

Function Calling

Prompt Response Writing SFT

Classification

Evaluation Rating

Freelancer Overview

I have hands-on experience as an AI training data consultant, where I evaluate and refine AI outputs to ensure quality, accuracy, and clarity, directly contributing to the development of ethical and high-performing language models. My background in cognitive science, paired with years of analyzing sales and performance data in fast-paced e-commerce and retail environments, has given me a strong foundation in data interpretation, attention to detail, and process optimization. I am adept at following complex guidelines, identifying inconsistencies or bias, and providing clear, actionable feedback to improve algorithmic outputs. My technical toolkit includes Microsoft Suite, Google Workspace, WordPress, and CRM platforms, and I am comfortable managing self-directed workloads to deliver high-quality results on time. I am eager to leverage my analytical skills and data annotation experience to support innovative AI projects.

Labeling Experience

Cross-Application AI Knowledge Retrieval Testing

Data Annotation TechTextRLHFRed Teaming

Contributed to an integrated workspace AI evaluation project focused on testing cross-application information retrieval and grounded response generation. Designed structured prompts to query connected data sources, evaluated outputs for factual accuracy, relevance, reasoning quality, and proper source grounding, and labeled responses for hallucinations and instruction adherence. Maintained strict compliance with detailed annotation guidelines and quality control standards, where continued project access and increased compensation were performance-based and contingent on accuracy. Contributed approximately 360 hours to the project series, consistently meeting precision and review benchmarks.

Contributed to an integrated workspace AI evaluation project focused on testing cross-application information retrieval and grounded response generation. Designed structured prompts to query connected data sources, evaluated outputs for factual accuracy, relevance, reasoning quality, and proper source grounding, and labeled responses for hallucinations and instruction adherence. Maintained strict compliance with detailed annotation guidelines and quality control standards, where continued project access and increased compensation were performance-based and contingent on accuracy. Contributed approximately 360 hours to the project series, consistently meeting precision and review benchmarks.

2025 - 2025

Video-to-Text Reasoning Validation Project

Data Annotation TechVideoQuestion AnsweringObject Detection

Contributed to a multimodal AI evaluation project focused on assessing video comprehension and contextual reasoning. Developed structured video scripts and generated AI chat interactions designed to test the model’s ability to extract and interpret information from video content. Created targeted question-answer pairs, validated outputs against ground-truth responses, and annotated results for accuracy, relevance, and reasoning quality. Completed 50+ video-based evaluation tasks over approximately 50 hours while adhering to detailed annotation guidelines and quality review standards.

Contributed to a multimodal AI evaluation project focused on assessing video comprehension and contextual reasoning. Developed structured video scripts and generated AI chat interactions designed to test the model’s ability to extract and interpret information from video content. Created targeted question-answer pairs, validated outputs against ground-truth responses, and annotated results for accuracy, relevance, and reasoning quality. Completed 50+ video-based evaluation tasks over approximately 50 hours while adhering to detailed annotation guidelines and quality review standards.

2025 - 2025

Research-Driven Prompt Engineering & Model Evaluation

Data Annotation TechTextClassificationRLHF

Contributed to an adversarial AI evaluation project focused on designing research-driven prompts to test model robustness and expose reasoning, factual, or instruction-following failures. Conducted extensive external research to develop prompts aligned with predefined stress-testing criteria, with the objective of systematically inducing performance breakdowns in one of two models under evaluation. Generated structured prompts, engaged models in controlled testing scenarios, and applied classification and preference-based scoring to assess response accuracy, coherence, and compliance with task requirements. Operated under strict evaluation guidelines and quality review standards to ensure precision, reproducibility, and adherence to defined testing objectives.

Contributed to an adversarial AI evaluation project focused on designing research-driven prompts to test model robustness and expose reasoning, factual, or instruction-following failures. Conducted extensive external research to develop prompts aligned with predefined stress-testing criteria, with the objective of systematically inducing performance breakdowns in one of two models under evaluation. Generated structured prompts, engaged models in controlled testing scenarios, and applied classification and preference-based scoring to assess response accuracy, coherence, and compliance with task requirements. Operated under strict evaluation guidelines and quality review standards to ensure precision, reproducibility, and adherence to defined testing objectives.

2024 - 2025

Comparative LLM Conversational Evaluation

Data Annotation TechAudioClassificationRLHF

Contributed to a conversational AI evaluation project focused on comparing the performance of two language models using identical prompts and dialogue conditions. Engaged each model in parallel conversations and performed structured evaluation of response quality, including coherence, contextual retention, reasoning, instruction adherence, and overall conversational effectiveness. Applied classification and preference-based ratings to assess relative performance and document strengths and weaknesses across outputs. Completed approximately 600 of comparative evaluation work while adhering to strict quality guidelines, detailed scoring rubrics, and review standards to ensure consistent, objective, and reproducible assessments.

Contributed to a conversational AI evaluation project focused on comparing the performance of two language models using identical prompts and dialogue conditions. Engaged each model in parallel conversations and performed structured evaluation of response quality, including coherence, contextual retention, reasoning, instruction adherence, and overall conversational effectiveness. Applied classification and preference-based ratings to assess relative performance and document strengths and weaknesses across outputs. Completed approximately 600 of comparative evaluation work while adhering to strict quality guidelines, detailed scoring rubrics, and review standards to ensure consistent, objective, and reproducible assessments.

2024 - 2025

Structured LLM Evaluation & Rubric Development

Data Annotation TechTextClassificationRLHF

Contributed to an AI conversation evaluation project focused on assessing factual accuracy, reasoning quality, and instruction adherence in model-generated responses. Reviewed completed AI-user dialogues across diverse subject areas, conducted external research to verify claims, and developed atomic, independently verifiable grading rubrics to systematically evaluate response quality. Applied structured scoring criteria to assess correctness, completeness, logical consistency, and policy compliance. Contributed approximately 700 hours while adhering to strict quality standards and review protocols to ensure objective, reproducible evaluations.

Contributed to an AI conversation evaluation project focused on assessing factual accuracy, reasoning quality, and instruction adherence in model-generated responses. Reviewed completed AI-user dialogues across diverse subject areas, conducted external research to verify claims, and developed atomic, independently verifiable grading rubrics to systematically evaluate response quality. Applied structured scoring criteria to assess correctness, completeness, logical consistency, and policy compliance. Contributed approximately 700 hours while adhering to strict quality standards and review protocols to ensure objective, reproducible evaluations.

2024 - 2025

Education

Work History