Call for applications is now open!
The iROSA group is seeking to hire enthusiastic researchers that want to work on highly interdisciplinary topics at the intersection of classical robotics (planning & control), robotic multi-sensorial perception, and machine learning methods.
iROSA seeks to identify structure in robotic systems, in the geometry of the perceived environment, and task specifications and use it for learning to solve complex long-horizon tasks with mobile manipulation robots in the non-stationary, ever-changing real world.
Offered topics involve:
- Geometric deep robot learning;
- Graph Neural Network representations for Robot Learning;
- (Active) Robotic Multisensory Perception for manipulation and interaction;
- Learning for (grounded) Mobile Manipulation coordination;
- Uncertainty-aware robot learning for (mobile) manipulation;
- Human behaviour understanding, and uncertainty quantification;
- Safe and Explainable physical Human-Robot Interactions (e.g., handovers, care tasks like wiping);
- Long-horizon reasoning, with a focus on combining Graph-based Reinforcement Learning and Planning.
Our policy is first-come first-served!