Toronto, Ontario, Canada
I am interested in contributing to endeavours where formal predictive models and advanced statistical techniques are used to recognize behavioural patterns, uncover insights, and answer challenging questions. I have a strong and broad background in sophisticated quantitative approaches to research and data analysis. I am particularly well versed in Bayesian approaches to both modeling and data analysis.
• Developed flexible and massively parallel scorecard development tools in Python. • Built and deployed an automotive loan early payout attrition predictive model to mitigate loan losses by identifying clients likely to churn and incentivizing desirable individuals to re-book with the bank. • Automated regulatory modeling processes necessary for expected credit loss provisioning. • Evaluated machine learning advances and demonstrated ways to incorporate innovative feature and model selection techniques into existing methodologies and processes. • Explored novel data sources and built data infrastructure to facilitate their consumption and use. • Researched explainable machine learning techniques necessary when operating in strict regulatory environments and used learnings to provide post-hoc explanations of black-box models and to fabricate prediction algorithms with robust performance that maintained intrinsic transparency.
• Developed predictive modeling infrastructure for employee flight risk prediction feature in Dayforce HCM software. • Leveraged machine learning techniques to create classifier models that predicted employee churn based on employee attributes and recorded behaviors. • Built automated time series data pipelines to verify data integrity, detect trends, and identify outliers. • Crafted a model tracking platform to quantify predictive performance across multiple models and metrics. • Provided analytical consulting on best practices and approaches for statistical analysis.
• Generated many predictive models forecasting client product use behavior (foreign currency exchange, global money transfer, etc.); models were utilized by customer marketing to target receptive clients and achieved lifts of up to 500% relative to baseline propensity. • Built and monitored novel predictive modeling infrastructure in SQL, SAS, and R, including a monthly data resource of 1,000+ handy predictor variables and a semi-Bayesian machine learning algorithm for rapid scorecard development; infrastructure enhancements decreased model development time by more than 50%. • Worked closely with partners across various lines of business to ensure that modeling project targets were aligned with business objectives and that model features were selected/engineered to reflect client behavior and demographics sensibly and transparently. • Proactively promoted and influenced adoption of analytical results, transferred in-depth knowledge of predictive models to business partners, and monitored operations to ensure models were properly used in customer communication.
• Generated many predictive models forecasting client acquisition behavior (mortgage, creditor insurance, etc.); models were utilized by customer marketing to target receptive clients and achieved lifts of up to 500% relative to baseline propensity. • Worked closely with partners across various lines of business to ensure that modeling project targets were aligned with business objectives and that model features were selected/engineered to reflect client behavior and demographics sensibly and transparently. • Created production quality automated scoring and tracking programs, exercising due diligence, rigor, and thoroughness in a timely fashion. • Reviewed results and communicated findings in understandable business terms. • Provided expert consultation on sales leads quality, articulating relevant findings as customer insights for publication.
• Provided campaign analyses - quantifying progress and conveying impacts along multiple performance metrics, such as contacts, sales, cross-sells, expected revenue, lift and ROI; providing insightful recommendations to improve the current design, avoid mistakes, and maximize efficiency. • Improved marketing campaign measurement methodologies by revamping the control group design to ensure proper measurement while enlarging the number of marketable clients available. • Implemented a method for multi-contact sales attribution using Shapley regression, enabling fair sales allocations across multiple campaigns for enhanced campaign effectiveness assessment. • Collaborated with business partners to ensure campaign designs were measurable, promoting constant communication to ensure awareness of available insights. • Conducted internal walk-through of campaign results and ensured insights were communicated to business partners. • Assimilated knowledge accumulated across different campaigns into a body of knowledge organized by line of business. • Absorbed marketing aptitude by being immersed in expansive projects such as customer lifecycle management optimization and do-not-solicit mitigation.
• Studied human cognitive processes including learning, categorization, recognition, and expertise, using computational cognitive modeling. • Designed, conducted and analyzed numerous experiments with human participants, including both lab-based and web-based experiments. • Developed and fit mathematical models to data to evaluate formal theories and make novel predictions.
• Studied human cognitive processes including learning, attention, perception, categorization, recognition, and decision-making, using computational cognitive modeling. • Designed, conducted, and analyzed numerous experiments with human participants. • Developed and fit mathematical models to data to evaluate formal theories and make novel predictions. • Coordinated the activities of multiple cognitive psychology labs, acting as the primary supervisor for many undergraduate research assistants. • Published multiple research articles and presented research findings at academic conferences.