Toronto, Ontario, Canada
Led Agentic Application initiatives within the IT Data & AI team, focusing on enterprise GenAI applications, evaluation, and business adoption. Built a GenAI PoC in Microsoft Fabric and advanced it toward UAT through stakeholder alignment and demos. Designed an LLM-as-a-judge regression evaluation framework with a human-in-the-loop dashboard, reducing manual review time by 78%. Also developed a Volume Analysis Agent for executive users by engineering the semantic layer, standardizing business terms, synonyms, and calculated metrics to improve natural-language-to-data interpretation.
Worked with the Enterprise Model Risk Management team on AI model validation, focusing on the assessment of an internal RAG application. Developed a validation framework covering the RAG pipeline, engineered benchmarking and sensitivity tests, and identified 20+ methodological, implementation, and documentation findings. Delivered actionable recommendations to improve model robustness and contributed to a comprehensive 70+ page validation report.
Supported Greater China Sales Data Analysis by automating data processing, KPI tracking, and reporting workflows. Revamped 15+ Python programs to streamline large-scale manual processes, reducing processing time by over 80% and lowering storage and computational costs by over 60%.