Sarah A.

Key Skills

Software

Internal/Proprietary Tooling

Top Subject Matter

Education / EdTech (Primary & ESL Learning)

Learning Science, Assessment & Curriculum Design

AI-Generated Content Evaluation & Prompt Design

Top Data Types

Text

Image

Top Task Types

Prompt + Response Writing (SFT)

RLHF

Evaluation/Rating

Freelancer Overview

I have directly transferable experience in AI training and data labeling through hands-on work with AI-generated content in educational products. I design prompt frameworks, create structured prompt–response pairs, and evaluate outputs against defined criteria to ensure accuracy, clarity, safety and age-appropriate tone. I iteratively refine prompts and guidelines to improve output quality and consistency, aligning closely with SFT and RLHF-style workflows. Alongside this, I have led quality assurance across large, structured content datasets, writing and applying taxonomies, guidelines, and editorial standards to make consistent decisions at scale. This includes reviewing thousands of assessment items and learning materials for correctness and clarity. I bring strong attention to detail, structured thinking, and the ability to apply evaluation frameworks consistently to produce high-quality training data for AI systems.

Labeling Experience

AI Image Evaluation & Prompt Optimisation

ImageEvaluation/Rating

Reviewed and evaluated AI-generated images for quality, accuracy, consistency, and age-appropriateness within an educational context. Assessed outputs against defined criteria, including visual clarity, alignment with prompts, and suitability for young learners. Identified common failure patterns and inconsistencies, providing structured feedback to improve output quality. Iteratively refined image prompts and developed guidelines to standardise outputs, effectively improving consistency and reducing error rates. This work involved classification of image quality and alignment, as well as applying consistent evaluation frameworks similar to image annotation workflows.

2021 - 2026

AI Output Evaluation & Prompt-Response Development

TextEvaluation/Rating

Worked extensively with AI-generated text outputs in an educational context, reviewing and refining responses for quality, accuracy, clarity, and age-appropriateness. Created and iterated on structured prompt-response pairs to support AI-driven features such as book generation, reading assessments, and teacher resources. Applied consistent evaluation criteria and editorial guidelines to assess outputs, identify errors, and improve alignment with pedagogical objectives. This included classifying outputs based on quality, correctness, and suitability for different learner levels. Contributed to iterative feedback loops that improved overall AI performance and consistency, closely aligning with supervised fine-tuning (SFT) and text annotation workflows.

2021 - 2026

Education

Work History