Chenggang Liu

About

A curious learner and a passionate explorer in the fields of AI, robotics, and automation

2025 – Present

Motional AD

Bridge the gap between research north star and reality, bringing the next-gen motion planning system to life. Nothing is too big, and nothing takes too long.

2023 – 2025

Meta Reality Lab Research

With the rise of GenAI, took a brief detour from motion intelligence to work on photorealistic Codec Avatars. The experience reshaped how I think about applying generative AI to motion planning.

2021 – 2023

Aurora Innovation

Worked on a parametric trajectory sampler for L4 self-driving cars. The challenge in sample efficiency led us to explore ML approaches for trajectory generation.

2015 – 2021

Uber ATG

Worked on trajectory generation, solving three major challenges: runtime performance; behavior formulation and tuning; and cost function learning.

2013 – 2015

Robotics Institute — Carnegie Mellon University

Worked on optimization-based motion planning and control for humanoid robots. Participated in the DARPA Robotics Challenge with Team WPI-CMU, leading the egress task. The team completed the task in under four minutes with a 100% success rate and placed 7th in the DRC final.

2011 – 2013

GE Aviation Systems Co. Ltd. and AVIAGE SYSTEMS

Worked on requirements and architecture development for a large commercial aircraft, building solid systems engineering and project management skills.

2008 – 2010

Robotics Institute — Carnegie Mellon University

Pioneered the use of an optimal trajectory library for motion planning and control. The trajectory optimization method I used later became one of the most efficient trajectory generation methods for self-driving cars.

Earlier

S3 Graphics

Worked on GPU logic design, amazed by the compute power within human reach, and started to wonder what the real bottleneck for AI really is. It sparked my interest in harnessing computational power to drive advanced AI.