Refined a time series model for predicting customer service call volumes, achieving a 30% relative reduction in error and enabling more efficient staffing to maintain service levels at reduced cost.
Explored large language models and topic generation techniques to summarize and categorize themes in complaint calls, providing actionable insights to optimize complaint handling processes and identify emerging trends.

Developing and monitoring end-to-end solutions on Canada's largest automotive marketplace, used Python and SQL on Azure, Databricks, and DataLake.
Built a Catboost classification model to predict approved car loan interest rates based on vehicle and customer data, achieving an average error below 1 and an R² over 0.85.
Defined key business metrics and developed data pipelines to quantify the impact of car ad boosting products, supporting marketing strategies and performance benchmarking.
Designed pipelines to build a model extracting vehicle features from images, supporting streamlined processing.

Built on prior research published in CHI 2023, co-authored during a MITACS Globalink Research Internship.
Tackled the challenges in understanding and evaluating Haptic Experience (HX), a subset of UX.
Aimed to create a tool that enhances the assessment of HX in product design and evaluation.
Designed and conducted user studies with over 600 participants across in-person and online formats, using a mix of qualitative (e.g., literature review, user interviews, expert feedback, and card sorting) and quantitative (e.g., factor analysis, measurement invariance testing) methods.
Identified five key dimensions of haptic interactions that shape the user experience, offering a deeper, actionable understanding of HX. Developed the first standardized tool for assessing HX, enabling streamlined, reliable evaluation of haptic experiences, improving product design processes.
Paper accepted to CHI 2025. Tool publicly available, facilitating standardized HX assessments in product design.