About me

I’m Zongwen Mu, a programmer, a retired Rubik’s cube player and a huge nerd. The only thing I succeeded in my life is failure.

I received my Master’s degree in Mechanical Engineering from Carnegie Mellon University and my Bachelor’s degree in Mechatronic Engineering from Zhejiang University, my research and academic work primarily focused on the perception field in robotics, specifically in computer vision and object detection.

I’m currently working as a perception engineer at AutoX.ai, focused on the perception part for L4 level autonomous driving, and I’m responsible for the development and maintainance of the traffic light detection pipeline.

My work at AutoX includes:

• Onboard code development: Refactor our onboard traffic light recognition pipeline, because the previous pipeline consists of a batch of complex code files and is not compatible with our latest perception engine, so I got to reorganize these modules and rebuild them in our workflow format. And this workflow is dramatically changed after we have deployed our time-sequential traffic light classification model.

• Deep learning model: Including the maintainance of our single frame detection models, and after the idea of multi-frame classification model was raised, I built a brand new traffic light classifier using Transformer to handle time-sequential related issues.