Susan Henry

Key Skills

Software

Appen

Data Annotation Tech

Labelbox

Other

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Image

Medical Dicom

Text

Video

Top Label Types

Action Recognition

Audio Recording

Classification

Diagnosis

Entity Ner Classification

Evaluation Rating

Fine Tuning

Object Detection

Prompt Response Writing SFT

Question Answering

RLHF

Text Generation

Text Summarization

Transcription

Freelancer Overview

I am an experienced AI QA reviewer and data annotation specialist with over six years of hands-on experience supporting large language model training and multimodal dataset validation. My work includes image-based question and answer review, answer correction, classification, rubric-based grading, and structured evaluation across text, audio, video, and image datasets. I have maintained 95%+ QA accuracy across high-volume review cycles and contributed to reducing annotation defects by 25–30% through systematic auditing and precise guideline adherence. I am particularly skilled at verifying alignment between visual content and textual responses, extracting exact text from images, and producing clear, standalone corrected answers. In addition to QA review, I have supported reinforcement learning workflows through pairwise comparison and response grading, contributed to a 15% improvement in transcription and language model processing accuracy, and validated scientific and technical content using strong analytical reasoning. My background in bioinformatics and computational biology strengthens my ability to detect subtle factual inconsistencies and reasoning gaps. As a native English writer, I bring strong grammar, clarity, and attention to detail to every task, ensuring consistent, high-quality labeling and evaluation outcomes.

Labeling Experience

Image Q&A QA Reviewer

AppenImageClassificationQuestion Answering

Reviewed image based question-and-answer items to verify alignment between visual content and associated textual responses. Assessed pre-labeled answers for factual accuracy, completeness, and instruction adherence, and applied corrected labels where necessary. Produced clear, standalone corrected answers supported by exact text extraction from images. Performed structured evaluation using defined QA rubrics to assess coherence, reasoning quality, and visual-text consistency. Identified errors including misinterpretation of image content, incomplete answers, and hallucinated details. Maintained consistent quality performance while meeting defined pacing targets and guideline updates.

Reviewed image based question-and-answer items to verify alignment between visual content and associated textual responses. Assessed pre-labeled answers for factual accuracy, completeness, and instruction adherence, and applied corrected labels where necessary. Produced clear, standalone corrected answers supported by exact text extraction from images. Performed structured evaluation using defined QA rubrics to assess coherence, reasoning quality, and visual-text consistency. Identified errors including misinterpretation of image content, incomplete answers, and hallucinated details. Maintained consistent quality performance while meeting defined pacing targets and guideline updates.

2025 - 2025

Multimodal LLM Evaluation and QA Annotation Specialist

Data Annotation TechTextText GenerationAction Recognition

Conducted structured annotation and quality auditing for large-scale multimodal AI training datasets, including text, image-based question and answer tasks, audio, and video data. Performed LLM response evaluation using rubric-based scoring frameworks to assess coherence, reasoning quality, factual accuracy, and instruction adherence. Completed pairwise comparisons and graded AI-generated outputs as part of reinforcement learning workflows. Handled audio transcription and speech validation tasks to support supervised model training. Performed scene-level video annotation and classification to improve downstream model performance. Maintained 95%+ QA accuracy while contributing to a 25–30% reduction in annotation defects through systematic guideline interpretation, consistency checks, and structured feedback. Worked within defined quality control benchmarks and pacing requirements to ensure data integrity and reliable model training outcomes.

Conducted structured annotation and quality auditing for large-scale multimodal AI training datasets, including text, image-based question and answer tasks, audio, and video data. Performed LLM response evaluation using rubric-based scoring frameworks to assess coherence, reasoning quality, factual accuracy, and instruction adherence. Completed pairwise comparisons and graded AI-generated outputs as part of reinforcement learning workflows. Handled audio transcription and speech validation tasks to support supervised model training. Performed scene-level video annotation and classification to improve downstream model performance. Maintained 95%+ QA accuracy while contributing to a 25–30% reduction in annotation defects through systematic guideline interpretation, consistency checks, and structured feedback. Worked within defined quality control benchmarks and pacing requirements to ensure data integrity and reliable model training outcomes.

2025 - 2025

Scientific Content Validation and Dataset Annotation for AI Training

OtherTextClassificationQuestion Answering

Supported AI model training by annotating and validating scientific and biomedical datasets used in language and predictive modeling systems. Reviewed AI-generated responses for factual accuracy, domain consistency, and logical coherence within biological and healthcare contexts. Performed structured response grading using predefined rubrics to assess reasoning quality and instruction adherence. Identified failure patterns in model outputs, including hallucinations, factual inconsistencies, and incomplete reasoning chains. Collaborated with research and AI teams to refine dataset labeling standards and improve model reliability. Contributed to a 15% improvement in transcription and language model processing accuracy through systematic validation and feedback cycles.

Supported AI model training by annotating and validating scientific and biomedical datasets used in language and predictive modeling systems. Reviewed AI-generated responses for factual accuracy, domain consistency, and logical coherence within biological and healthcare contexts. Performed structured response grading using predefined rubrics to assess reasoning quality and instruction adherence. Identified failure patterns in model outputs, including hallucinations, factual inconsistencies, and incomplete reasoning chains. Collaborated with research and AI teams to refine dataset labeling standards and improve model reliability. Contributed to a 15% improvement in transcription and language model processing accuracy through systematic validation and feedback cycles.

2023 - 2024

Education

Work History