I am a research scientist working at the intersection of robotics, machine learning, and computer vision to equip embodied agents with greater understanding of the world. I'm particularly excited about building systems that generalize β from perception to decision making β across the open-ended complexity of the real world.
Most recently, this has meant building foundation models for robotics: I co-led the release of VLA Foundry, a unified open-source framework for training VLAs at scale, and was a primary contributor to the Large Behavior Models project. I am especially interested in recipes for zero-shot generalization, multimodal VLAs, offline evaluation techniques.
I have a PhD in Aeronautics and Astronautics from MIT, where I worked with the Robust Robotics Group at CSAIL advised by Dr. Nicholas Roy, where I spent a lot of time trying to make things fly fast and not hit things. Before that, I received my B.S. in Mechanical Engineering from UC San Diego.
Email: the.katherine.liu [at] gmail [dot] com
Recent News
- June 2026: I gave a talk, βPre-training for manipulation: Empirical studies and open tools for the science of robot pre-training,β at the VLA Pipelines workshop at ICRA 2026. (details)
- April 2026: We released VLA Foundry: A Unified Framework for Training Vision-Language-Action Models. (details)
- March 2026: We released a sneak peek of what we've been working on in policy learning. (details)
- July 2025: We are sharing our findings in our report: A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation.
- April 2025: CodeDiffuser: Attention-Enhanced Diffusion Policy via VLM-Generated Code for Instruction Ambiguity accepted to RSS 2025 -- see you in LA!
- January 2025: OmniShape: Zero-Shot Multi-Hypothesis Shape and Pose Estimation in the Real World accepted to ICRA 2025. See you in Atlanta!
- June 2022: I joined the Toyota Research Institute as a Machine Learning Research Scientist.
Recent Publications
CodeDiffuser: Attention-Enhanced Diffusion Policy via VLM-Generated Code for Instruction Ambiguity
Guang Yin Yin, Yitong Li, Yixuan Wang, Dale Mcconachie, Paarth Shah, Kunimatsu Hashimoto, Huan Zhang, Katherine Liu, Yunzhu Li RSS 2025 [PDF]