Simon Kahenya

Annotated and evaluated AI model responses for mathematics and computational reasoning tasks. Designed validation checks and evaluation rubrics for STEM-related datasets. Collaborated with technical leads to test and refine model interpretability and logical reasoning outputs.

Design and evaluate high-quality prompts for AI reasoning, mathematical problem-solving, and Python coding tasks. Conduct in-depth LLM testing and refinement across mathematics, logic, and computational tasks. Collaborate with cross-functional teams on model evaluation strategies to enhance reasoning consistency. Train evaluators and oversee alignment-based model training ensuring adherence to prompt quality frameworks. Integrate SciCode tools for model testing, dataset validation, and analysis of AI reasoning structures.

Supervised data quality assurance for mathematics domain prompts used in LLM fine-tuning. Led prompt evaluation teams focused on ensuring clarity, accuracy, and reasoning depth in datasets. Contributed to development of metrics for assessing LLM mathematical reasoning reliability.