Yanlong Zhao

Research Assistant @Illinois Institute Of Technology

Chicago, IL, US

Developed scalable, reproducible machine learning systems for modeling biological and biochemical networks using graph neural networks (GNNs),delivering reusable, efficient ML infrastructure that improves prediction accuracy while reducing computing and data bottlenecks.• Designed a configuration-driven machine learning pipeline with dataloaders, model registries, and configuration files in PyTorch/DGL, whichallowed for large-scale benchmarking of machine learning projects such as DREAM-GNN, MuSH• Built a robust evaluation framework with 10-fold CV, cold-start splits, and sparsity stress tests, supported by full metadata logging for reproduciblereruns and standardized validation of GNN variants.• Optimized preprocessing pipelines for sparse graphs and biological sequences using NumPy vectorization and parallel I/O, reducing data preptime by 30 %+ and speeding up model iteration.• Scaled distributed GPU training on Lambda GPU Cloud using PyTorch DDP and Dockerized environments, implementing automated checkpointingfor fault-tolerant, resumable large-scale experiments.

Shandong University

Bachelor's degree, Automation

The University of North Carolina at Chapel Hill

Doctor of Philosophy, Data Science

University of Rochester

Master of Science - MS, Electrical and Computer Engineering