Subing Xiang

Recycle and Reuse Mobile Application Design, U of T

• Led a team of four to select the second-hand trading project for researching, designing, developing, analyzing and testing.
• Created and implemented competitive analysis, surveys, and interviews to collect qualitative and quantitative data to analyze user's preferences.
• Designed and developed wireframes and high-fidelity prototype by using Figma and PS and conducted usability tests to conclude findings and recommendations to increase user experience in the next phase.
• Presented the prototype that improved user satisfaction by 80% to the clients and teaching team, and received the highest score for the final report and presentation.

Polynomial Local Minimizer Project, U of T (Java)

• Built the framework (classes and objects) for the polynomial local minimizer by using Java, and the Steepest Descent Algorithm was applied
• Utilized Armijo and Golden section line searches to test the learning rate and performance

Traffic Sign Recognition Project, U of T (Python)

• Conducted multiclassification deep learning project based on PyTorch framework.
• Collected data from online databases about 11 common traffic sign images.
• Applied end-to-end CNN model and auto-encoder model to train the model and do the inference, and tried transfer learning based on pre-trained VGG model.
• As the result, the best performing model reached 97.7% accuracy on the testing set.