Ed Chen

About Ed Chen

Chemistry × Electrical Engineering × Programming

Experience

1. Machine Learning and Circuit Design Intern

   Synopsys Inc

   May 2025 — Present · Austin, TX, US

   Developed a Circuit Foundation Model (CFM) for VLSI circuits using Recurrent Hypergraph Neural Networks (RHGN), representing circuits as hypergraphs to generate embeddings for devices, wires, and terminals based on their connectivity.• Designed and implemented advanced message-passing (\"hopping\") algorithms to recursively capture and aggregate connection information across circuit components, enabling robust circuit representation learning.• Built modular utilities—including a TrainingRecorder for loss tracking, ConfigManager for centralized hyperparameter management, and ResourceManager for real-time memory/CPU monitoring—to streamline and maintain complex training workflows.• Automated the visualization of neural network weights and biases at each training epoch, producing interpretable heatmaps and GIFs to enhance model transparency and facilitate diagnostics.• Engineered custom loss functions and evaluation metrics (including confusion matrices, precision, recall, and F1-score) to effectively address severe class imbalance and improve model performance assessment.• Developed scalable data ingestion and preprocessing pipelines capable of handling large-scale circuit datasets, supporting flexible batch training and efficient experimentation.

Education

• UC San Diego

  Environmental Engineering

  2014 — 2014

• National Taiwan University

  Master's degree, Physical Chemistry

  2018 — 2020

• HSNU - The Affiliated Senior High School of National Taiwan Normal University

  High School Diploma, Math and Science Gifted Class

• National Tsing Hua University

  Bachelor's degree, Chemistry

  2012 — 2016

• Massachusetts Institute of Technology

  Doctor of Philosophy - PhD, Electrical Engineering and Computer Science