Featured Projects
-
IEEE RA-P 2026
MOMO
Adapt a robot's skill by touching it, talking to it, or clicking on it
-
🏆 ICRA 2026 Spotlight
CLASP
A robot that understands what you want — and admits when it doesn't know how
-
🏆 ICRA 2026 Spotlight
IROSA
Adapt a robot's behavior just by talking to it
-
IEEE RA-L 2025
LOCI
Teach a skill once, then refine it interactively — without forgetting the original
Research Focus
I develop methods that enable robots to learn new skills interactively and adapt their behavior through natural language. A person shows the robot a movement a few times, and the robot turns those demonstrations into a skill it can adjust and reuse. My work combines probabilistic machine learning with foundation models — progressing from learning by physical demonstration, to language-guided motion adaptation, toward fully autonomous robot agents.
Recent News
- Aug 2026 MOMO published in IEEE RA-P — with a new official video
- Aug 2026 Code for LOCI and IROSA is now runnable as Code Ocean capsules
- Jun 2026 Received two Next-Gen Spotlight Awards at ICRA 2026 in Vienna for CLASP and IROSA
- Apr 2026 New project MOMO: Adapt a robot’s skill by touching it, talking to it, or clicking on it
- Apr 2026 New framework CLASP: natural language-driven robot skill selection and composition
- Mar 2026 Paper published in IEEE RA-L: “IROSA: Interactive Robot Skill Adaptation using Natural Language”
- Aug 2025 Attended ELLIS Cambridge & OxML Oxford summer schools on probabilistic ML and representation learning
- Jun 2025 Live demo of LLM + KMP robot control at Automatica 2025 in Munich
- May 2025 Co-organized the “Queer in Robotics: Building a Community and Generating Inclusive Guidelines” workshop at ICRA 2025
- Apr 2025 Co-authored paper at ICRA 2025: “Grounding Embodied Question-Answering with State Summaries”
- Feb 2025 Paper published in IEEE RA-L: “Interactive Incremental Learning of Generalizable Skills”
- Nov 2024 Spotlight presentation at the CoRoboLearn workshop, CoRL 2024