Ziang Xu Email & Phone Number

Atlanta, Georgia, United States

- Designed and developed a Scheduling as a Service algorithm that achieved customer production of 1M+ items/week sourcing from 1K+ farmhouses for ~3 years in advance.- Developed CI/CD pipeline on Azure to streamline testing and productization for the data science team, project managers, and engineering team, leading to an increase in feature development turnaround time by ~30%.- Collaborated with the engineering team to refactor 120K+ lines of code into a more readable, testable, and manageable version.- Implemented 200+ unit and integration tests, leading to significantly enhanced deployment stability and robustness.- Implemented complex business logic in the IP model by the big-M technique. - Improved previously elusive schedules through IP, resulting in a 25% enhancement in schedule quality measured against custom metrics.

Ann Arbor, Michigan, United States

- Optimized a BEM (building energy model) simulator to gain deeper insights into building energy behavior.- Devised a block-gradient descent algorithm, and proved its convergence and bounded convergence rate.- Implemented the algorithm from scratch as the existing libraries cannot be incorporated with the BEM simulator; achieved 30% accuracy improvement compared to commonly used algorithms; results were published in 2 papers. - Architected a machine learning solution incorporating Linear Regression and Random Forest to predict torque and rotation speed for Ford electric engines, achieving a 30% accuracy improvement compared to the previous year’s model.

Ann Arbor, Michigan, United States

- Hosted remote office hours and prepared online tutorials to teach students to implement their ideas in data analytics with Python and R.

Ann Arbor, Michigan, United States

- Applied machine learning models in supply chain demand prediction - Collaborated with senior personnel in EDA (exploratory data analysis), feature engineering, and model selection- Improved accuracy by 25% via fine-tuned Random Forest- Identified the relationships between seasonality, variance difference, and forecasting model selection

Ann Arbor, Michigan, United States

- Increased 20% accuracy within 1 month in eye surgery postoperative diagnosis by implementing VAE-GAN to resolve the problem of insufficient data. - Collaborated with data scientists in data collection, cleaning, visualization, experiments design, and label selection; reduced loss value of product recommendation model by 15% over 3 months via fine hyperparameter tuning.

Master Of Science - Ms, Industry And Operations Engineering

Bachelor'S Degree, Industrial Engineering