Elias Rowan

Evaluated AI-generated data science and machine learning responses for analytical correctness, statistical rigor, code quality, and instruction adherence. Designed challenging training tasks and expert reference solutions across statistical modeling, machine learning, Python, SQL, and applied data analysis. Performed step-by-step reasoning review to identify flawed assumptions, methodological errors, ambiguous conclusions, and incorrect analytical code. Provided concise corrective guidance and technically sound explanations of appropriate methods, assumptions, metrics, and interpretations. Created and validated high-quality training examples and compared alternative approaches to select the most accurate and complete answer. Worked independently using detailed evaluation rubrics and iterative reviewer feedback while maintaining consistent quality across project cycles.

Reviewed model outputs, training data, and analytical solutions across machine learning, statistics, and applied data science workflows. Used Python and SQL for exploratory analysis, data validation, statistical testing, feature analysis, and model-performance assessment. Evaluated regression, classification, and predictive modeling approaches for sound methodology, appropriate metrics, and reliable interpretation. Ranked and corrected model responses, documented failure cases, and supported structured human-feedback workflows for AI improvement. Identified inconsistencies, reasoning gaps, and edge cases and translated findings into clear technical feedback, while maintaining quality against detailed evaluation criteria and incorporating iterative reviewer guidance. Produced clear rationales explaining why specific methods or conclusions were correct or incorrect.

Evaluated machine learning and AI outputs for factual accuracy, reasoning quality, consistency, and analytical correctness. Created and reviewed prompts, reference responses, training examples, and realistic data science evaluation scenarios. Applied statistical and machine learning methods to assess model behavior, investigate errors, and validate analytical conclusions. Reviewed Python-based workflows for coding issues, data leakage, inappropriate methods, and incorrect interpretations. Performed structured quality reviews of labeled data and model outputs, flagged ambiguous cases, and documented corrective guidance. Worked within detailed task instructions and quality standards to deliver consistent, independently reviewed project outputs.