VLA · imitation learning · robot learning
Vision-language-action models let a robot take a camera image and a text instruction and output actions. I help students and developers understand how they work and how to experiment with them.
Sessions combine the machine learning concepts with practical work on real hardware or in simulation.
Vision-language models, transformers, policies and how actions are represented.
Collecting demonstrations, training policies and evaluating them.
Python, PyTorch, Hugging Face LeRobot and ROS 2 basics.
Low-cost robot arms, cameras and the control loop that connects them.
Using language models to plan tasks and call robot skills.
A vision-language-action model takes camera images and a language instruction and predicts robot actions, in a single network.
Basic Python is required. Machine learning and robotics experience helps, and we adjust the pace to your level.
No. We can start in simulation or with a low-cost arm, and I will advise on options.
Fill out the form with the topic and your preferred time, and I’ll get back to you.
Request Tutoring →