Yingjue Zhou Email & Phone Number

Pleasanton, CA, US

Ventura, CA, US

• Develop Time Series models to forecast peak hours of high quality advertising opportunities. Increase spending on peak hours to enhance advertising performance.
• Analytics to identify a root cause of under spending due to suboptimal advertising campaign settings. Forecast the potential spend and interactively prompts customer to optimize the setting.
• Analytics to identify a problem in bidding system causing under spending. Improve the bidding algorithm by forecasting advertisers' bidding prices and prioritizing high bidders.
• Train regression models to forecast advertisers' daily spend given a bidding price factor. Design and implement control algorithms to dynamically align advertisers' spending to the daily target.
• A/B testing to validate Neural Network models to enhance click rate and conversion rate. Orchestrate the step-by-step rollout of the models across the platform.

New York, NY, US

• Conduct the Targeted Rebate Program, a customized email marketing campaign. Train classification models to identify customers having a better chance to shop with rebate.
• Demand forecast algorithm for fashion products – generate initial demand forecast based on similarity, and then periodically update with new selling data by Bayesian Inference.
• A new restocking algorithm – forecasting the demand and per-item margin, build a MILP models to maximize the expected margin minus combined shipping cost.
• A new inventory strategy – aggregate slow sale products from multiple fulfillment centers to one to reduce logistics cost significantly with little impact on customer satisfaction.
• A new fashion product allocation strategy – train regression models to forecast the fashion product sales percentage adjacent to the fulfillment centers and allocate accordingly.

Chicago, IL, US

• Size profiling optimization– design and implement a Genetic Algorithm framework to find the optimal assortment of different sizes to put into packs, and the optimal quantity of packs to order.
• Delivery frequency optimization – design and implement a simulation based framework to find the optimal weekly schedule for a fleet to provide home service, embedded with a vehicle routing algorithm with time and.

San Diego, CA, US

• Design and implement a fast heuristics algorithm (neighbor search, multiple objective optimizations) to optimize the relocation of Fire/EMS vehicles to provide coverage to neighborhoods.
• Design and implement a vehicle routing algorithm to fulfill medical transportation requests with time and capacity constraints

Stillwater, OK, US

• Available-to-Promise Decision Support System for Inventory Allocation Problem– Translate the customer’ requirements into mathematical models. Find the optimal cutting patterns for metal tubes in inventory to meet.
• Economic analysis on electricity markets with renewable penetration– Study the impacts of Renewable Portfolio Standard regulation and the associated Tradable Green Certificate transaction.– Study the cooperation.