Harrison Zhao

Meta Superintelligence Lab - RealTime Al (current) MRS (Meta Recommendation Systems): Built and optimized retrieval and ranking models across FB, IG, and Threads, spanning graph learning, contrastive learning, model architecture, and large-scale data optimization.

Use machine learning, optimization and data analytics to solve challenging problems in capacity planning/management for Azure.

Developed large-scale machine learning and optimization models that help plant breeders to make better decisions. My works involved building data pipeline, developing predictive/prescriptive model and application deployment. Tech Environment: Python, R, SQL, PyTorch, Keras, TensorFlow, LightGBM, XGBoost, CPLEX, AWS, Git, Domino Data Lab

Worked on a project using natural language queries to control a drone to fly in the desired trajectory. Built a recurrent neural network with the self-attention mechanism for language embedding.

Developed a navigation app with smart notifications for subway service changes. • Built a real-time rerouting system based on Google Map API and MTA service changes notifications. • Implemented natural language processing to analyze subway change service notifications and convert it to machine readable real-time data. • The app will passively notify users based on their regular commutes.

Technical Courses: Applied Machine Learning, Modeling under Uncertainty, Optimization Methods, Natural Language Processing, Computer Vision