Parsa Farshadfar

Senior Data Scientist, AI/ML Model Validation, Risk Dept.

• Independently led comprehensive validation projects on machine learning models used by TD Bank and TD Insurance departments across Canada and the US, ensuring high standards of analytical rigor and methodological soundness.
• Oversaw the development of machine learning models, focusing on enhancing the conceptual frameworks, assumptions, and data integrity on MS Azure Cloud platforms.
• Mentored interns, providing comprehensive training and guidance on technical skills.
• Led technical interviews to evaluate candidates' ML and data analysis proficiency.
• Integrated cutting-edge Large Language Model (LLM) technologies to drive continuous improvement in analytics capabilities.

Data Scientist, AI/ML Model Validation, Risk Dept.

• Conducted over 10 in-depth validations of ML/DL models developed by various TD departments across Canada and the US.
• Performed model redevelopment for the ML models, assessing the conceptual soundness of modeling methodologies, assumptions, and the quality of large-scale data on MS Azure.
• Enhanced the validation process by 25% by integrating advanced libraries and improving the time complexity of repositories on both Virtual Machines and Databricks, resulting in significant operational efficiency gains.
• Facilitated cross-departmental collaborations to align ML strategies with business objectives by developing a web application for automatic bias-fairness analysis of ML models, tailored for TD's Compliance and Model Validation teams; the application received the Best App Award in TD Risk Management in 2023 for its innovation and high performance.

Machine Learning Researcher

• Developed the Spiking Convolutional Neural Network (SCNN), an energy-efficient deep learning method for automated seizure detection, leveraging spatio-temporal features from multichannel EEG time-series data to achieve state-of-the-art precision and recall.
• Reduced computational costs by 40%, making the AI process ideal for low-power portable medical devices.

Teaching Assistant

• Design and Analysis of Algorithms Course
• Object-Oriented Programming Course
• Data Structures Course

Data Scientist

• Leveraged deep learning and machine learning techniques to enhance user experience and efficiency, analyzing datasets exceeding 10 TB.
• Boosted click-through rates (CTR) of Ads by 4% and enhanced advertiser revenue, achieving a 3% improvement in Return on Ad Spend (ROAS) through data-driven strategies and precision-targeted advertising initiatives.
• Transformed over 50 large-scale SQL datasets for analysis, employing various ETL techniques to ensure data integrity and accessibility.
• Simplified complex data analysis for non-technical audiences, creating over 7 interactive dashboards and visualizations that facilitated decision-making.