Subject Matter Expert – AI Training & Evaluation Consultant
Authored high-complexity gold-standard solutions and instructional datasets for Physics and Calculus using step-by-step reasoning to train and fine-tune language models. Critically audited and evaluated hundreds of AI-generated responses in advanced STEM topics, performing fact-checking and providing qualitative feedback on scientific and mathematical accuracy. Developed and applied adversarial and localization scenarios for red team evaluation and Swahili-language STEM content adaptation. • Wrote and synthesized explanations that strengthen model transparency in scientific methodology. • Deconstructed multi-variable problems into structured logical prompts for enhanced prompt engineering. • Audited over 500 AI responses and corrected hallucinations, technical errors, and formatting inconsistencies. • Evaluated, ranked, and provided feedback on model outputs, including localization of math word problems.