Dung-Yi Wu

About Dung-Yi Wu

A motivated and collaborative scientist with: 6+ years experience in data…

Experience

1. Research Engineer

   GE Aerospace

   Nov 2024 — Present · Niskayuna, NY, US

   Probabilistic Design and Materials informatics TeamBuilt machine learning models (deep neural networks (ANN, Graph Neural Network), tree-based models (XGBoost) and Gaussian-based models) to predict material properties, providing actionable insights to support strategic manufacturing decisionsDeveloped automated material screening pipelines from scratch using inverse models (e.g, generative AI, Bayesian optimization, and genetic algorithms) with Python (PyTorch, scikit-learn) to accelerating the development of new materials, improving 50+% ROIConducted root cause analysis, including feature engineering (implementation of pertained LLMs e.g. Bert models, Machine learning forced field - CHGNet, MACE) and sensitivity testing (Sobol & SHAP), to diagnose mechanisms behind property predictions and enhance model accuracy Constructed materials properties database using python and MySQL for catalysis, ceramics, and polymers for aerospace applications and sustainability purposesCollaborated with business and manufacturing stakeholders to communicate and interpret data insights to aid the strategic decisions

Education

• Cornell University

  Master of Science - MS, Materials Engineering

  2017 — 2019

• Johns Hopkins Whiting School of Engineering

  Ph.D., Materials Science and Engineering

  2019

• National Taiwan University of Science and Technology

  Bachelor of Science - BS, Material Science & Engineering

  2012 — 2016