I'm Weihao Sun.

Thesis Project: AI Recommendation for Lane-Changing Problem Applying Cyber-Physical Human System (CPHS) Framework.
Advisor: Dr. Andreas Malikopoulos

Coure Courses: Model Based Systems Engineering, Project Management, Optimal Control and Decision Theory, Engineering Stochastic Process, etc.

Major in Statistics, Minor in Data Science.
Core Courses: Introduction to Probability (A+), Statistical Inference for Data Science (A), Machine Learning and Data Mining (A), Methods for Statistical Learning (A), Statistical Modelling for Data Science (A)

- Analyzed purchase data for over 50 construction projects, identifying trends and patterns to optimize procurement processes, resulting in a 5% reduction in costs.
- Reviewed and verified monthly purchase detail calculations totalling over ¥200 million to ensure accuracy and compliance.
- Developed and maintained data dashboards, increasing reporting efficiency by over 30% and supporting decision-making.
- Collaborated with cross-functional teams to simplify the data collection and reporting pipeline, reducing processing time by over 25%.

- Developed, tested, and maintained an internally used automatic bill calculation system.
- Updated and maintained the company database; Analyzed the database structure; designed and modified the structure to cater to different use cases.
- Designed the Microsoft Power Automate pipeline to Use Excel Online to write Office Script to automatically fill in charging rates, amounts, and dates into the daily form.
- Matched and extracted essential information such as tracking numbers and prices from large, noisy data provided by clients for usage by other departments, which improved the working efficiency by over 60%.

- Physics Informed Neural Network for Battery Modeling Project.
- Solved partial differential equations to improve the current battery models with Physics Informed Neural Network (PINN)
- Conducted literature survey on PINN, SPM, and Lithium batteries; Collected, cleansed data, prepared data for data visualization, and applied exploratory analysis.
- Developed the Physics Informal Neural Network from scratch, fitted the model to existing data, and analyzed the PINN model.

Tutoring Statistics, Math, Computer Science, and Data Science courses at UBC and SFU.

- Aimed to enhance AI-driven recommendations for vehicle lane adherence and switching in autonomous driving systems by incorporating human behavior dynamics, improving decision-making reliability in real-world scenarios.
- Conducted a comprehensive literature review on CPHS frameworks, including mathematical and statistical derivations and proofs of lemmas, methodologies to adapt the framework for diverse application areas.
- Refined and extended the CPHS framework by introducing new mathematical formulations and functional components tailored to address the complexities of lane recommendation problems in mixed human-AI environments.
- Currently preparing for model validation and simulation using state-of-the-art autonomous driving platforms (e.g., CARLA), and optimizing AI recommendations through reinforcement learning techniques such as Q-learning.

- Supervised by Dr. Clifford Whitcomb and Dr. Yiye Zhang at Cornell University.
- Aimed to develop a clinical decision support tool for personalized mental health interventions in patients experiencing pregnancy loss, for the doctors to make decisions on predicted risk levels.
- Conducted literature review on methodologies to find factors influencing risk levels in post-pregnancy complications and predictive models for risk assessment.
- Currently working on generateing data visualization, summary, and analysis results using PRAMS healthcare data.

- Collaborated with Borealis AI, a research institute of RBC.
- Applied and evaluated advanced machine learning models, including Regularized Regression, Neural Networks, Time Series models, etc., to predict local emergency department demands.
- Performed in-depth feature creation by appending climate, holiday, and demography data to enhance analysis.
- Kept identifying opportunities for adaption and refinement, and seeking collaboration with local hospitals or public health agencies.

- Performed early-stage raw data collection via public channels, including National Statistics Bureau, FRED, etc.
- Applied Naive Bayesian analysis, KNN, and SVM strategies for model construction and autonomous testing.
- Applied classical statistical methods, including confidence and prediction intervals and their calibration to evaluate models.
- Adopted transformers including SVD and PCA for model improvements.

- Implemented and evaluated various machine learning models (k-nearest neighbor, decision tree, support vector machine, and logistic regression) to classify animal specifies based on physical features.
- Developed a reproducible pipeline for the classification problem, and integrated the pipeline using Docker.
- Developed, maintained, and published Python packages encapsulating commonly used functions in the pipeline to enhance reproducibility.

- Developed an full stack application from scratch using Django framework and SQLite.
- Implemented a database modelling functions on services, employment management, and amusement facilities.
- Performed database structure analysis, and normalized the relations among tables to reduce redundancy.
- Designed the user interface of the application with the Bootstrap template.

- Developed in the progress of both console-based and GUI-based versions and maintained simple transitions.
- A turn-based RPG where players engage in battles using a selection of spells, each with unique attack values, against progressively challenging enemies, enhancing the immersive Harry Potter-themed experience.
- Implemented a dynamic archive management system allowing players to choose from various wizards, each with distinct HP and ATK stats, adding strategic depth to battles and ensuring a varied gaming experience across different sessions.
- Included a well-designed interface, stories, characters and portraits.