Shijia Hu

Title: Shijia Hu and Machine-Learning-Based IC Design Planning.

Introduction

Shijia Hu is an inventor based in Santa Clara, California, and is associated with NVIDIA Corporation. His work includes a patented technique for planning integrated circuit (IC) layouts using machine-learning-based congestion prediction.

Latest Patents

Hu is credited with 1 patent: “Bounding Area Planning Using a Congestion Prediction Model.” The invention uses placement data for cells in an IC design to generate a graph, which is then supplied to a machine-learning model. Trained to predict congestion levels in bounding areas, the model produces congestion data that can guide the selection of cells for placement in the corresponding region. This approach connects congestion prediction with practical decisions in IC design planning.

Career Highlights

Hu’s patent reflects work at the intersection of machine learning and semiconductor design. By using graph-based representations of cell placement, the technique aims to help identify congestion in specific design regions and inform cell-placement decisions.

Collaborations

The supplied information identifies Tian Yang as Hu’s coworker. Both Hu and Yang are associated with NVIDIA Corporation.

Conclusion

Shijia Hu’s patent applies machine-learning-based congestion prediction to IC design planning, linking placement data and graph models to decisions about where cells should be installed.