Sarah Odunayo Oladunjoye

I’m an AI Prompt Engineer and LLM Evaluator with 4+ years of experience supporting AI training through prompt development, structured evaluation, and dataset validation. I focus on improving model outputs with quality-focused decision making across Generative AI / NLP initiatives.

In my work, I develop and test prompts for summarization, classification, reasoning, extraction, and instruction-following. I also rank and compare AI-generated responses for accuracy, relevance, clarity, and adherence to project requirements, including RLHF-style preference ranking workflows.

I’m especially strong in rubric-based evaluation, factuality assessment, hallucination detection, and safety review. I apply source verification and document findings to identify unsupported claims, biased content, unsafe outputs, and instruction-following failures—helping teams protect dataset integrity and evaluation standards.

Across remote projects, I’ve contributed to large-scale annotation and QA, from reviewing 1,000+ outputs to validating 4,000+ text, image, and audio samples. I build reusable prompt libraries and LLM evaluation frameworks to support consistent, repeatable training and grading outcomes.