• Employed Python for data manipulation and preprocessing tasks, including data reading, type conversion, filtering, handling missing values, column manipulation, and data transformation.• Utilized Python to develop monthly sales visualizations, which included identifying top-selling products, analyzing highest profit margins, and assessing sales volume. Visualized outputted data using Power BI. Sales increased by 400% after the second month.• Retrieved ratings and locations of all vape shops in Ontario using Python and Google Cloud Platform ’Place’ API. Exploited PowerBI to plot ratings as bubbles on map, facilitating analysis of competitiveness and potential improvements. Increase Google Maps user views and route inquiries by 2 times.• Developed and managed an online vape store [n1vape.ca] using Shopify, implementing custom themes, optimizing the user experience, integrating payment gateways, and leveraging analytics to monitor customer behavior and conversion rates, leading to data-driven improvements in store efficiency and profitability.

• Optimized the Work Management System using C++ and MYSQL, effectively addressing and resolving user conflicts that may arise during weekly report submissions.• Deployed the Vehicle Color Classification Model within Docker containers on Cloud. Integrated the model with an AI training platform engine, actively monitored Docker logs, validated classification accuracy through front-end website.• Improved detection speed by 30% when applied to subway turnstile gates by utilizing Python and SQL to craft Face Detection and Alignment algorithm for the 3D Face Recognition Project. The model was fine-tuned by using various backbones.• Performed A/B testing and fine-tuned face detector and alignment models with the results. The model performance surpassed competitors, achieving 90.43% accuracy, 1.19% error rate, 8.48% omission rate.• Published 2 patents: Invention and Design, for employing Python and SQL in initiating the Palm Keypoint Detection algorithm for Palm Vein Recognition Project. Achieved 100% accuracy in capturing the center of palm and identifying left and right hands.

• Applied Python to construct the Data Processing Algorithm for Moving Object Tracking. Implemented DBSCAN Clustering for data cleaning. Filtered out 80% of the noises, enabling efficient tracking within a 200ms timeframe.• Initiated the project management for the Driver Monitor System Project by outlining the schedule and deliverables. Utilized Microsoft Project to create a Work Breakdown Structure and identified potential risks.