Hi, this is Xinyan (Shane) Yan. I am a software engineer at Aurora Innovation, working on solving motion planning for self-driving cars through machine learning. In July 2020, I completed my PhD in Robotics at the Georgia Tech Robot Learning Lab where I was advised by Byron Boots.

I am interested in developing theories, algorithms, and good practices for solving real-world sequential decision making problems. My current focus is imitation learning that leverages human demonstrations and interventions at scale to circumvent the curse of dimensionality and the difficulties in modeling while ensuring safety. Previously, I worked on a wide range of robotic problems, including SLAM, motion planning, system identification, and stochastic optimal control.

I was awarded the IJRR Paper of the Year 2018 and Finalist for Best Systems Paper, RSS 2018.

During my PhD study, I spent a summer and a fall at Google Brain where I worked with Krzysztof Choromanski, Vikas Sindhwani, Aleksandra Faust, and James Davidson. I also interned at the Vision and Sensor Group at DAQRI working with Jason Chen-Chi Chu and Wenyi Zhao, and at the Perception Team at Honda Research Institute working with David Ilstrup and Aniket Murarka.

Before coming to Georgia Tech, I graduated from Shanghai Jiao Tong University with a B.S. in Information Security Engineering.