PhD Position Physics-informed Foundation Models for Robotics

Develop physics-informed techniques for the next generation of robotic foundation models!

We are looking for a PhD candidate to join the Physical Intelligence Lab at TU Delft and contribute to the development of physics-informed foundation models for robotic manipulation. The position is part of the European project GRAIL, which aims to extend the paradigm of foundation models to systems interacting with the physical world.

In big tech, a race is underway to collect as much robotic data as possible, with the goal of training the ChatGPT of robotics physical intelligence. But physical data are much harder to get than language examples.

The PhD candidate will help solve the robotic data gap by developing learning architectures that integrate physical structure from mechanics and dynamical systems into modern machine learning frameworks. This way, the models will not have to learn physics from data every time and will be able to focus only on what is actually new.

The work will involve developing representations that enable generalization across tasks, environments, and robotic platforms, with particular attention to deformable media and compliant manipulation.

The candidate will be supervised by Dr. Della Santina and work within a collaborative research environment spanning control theory, machine learning, and (soft) robotics, with access to experimental platforms and collaborations across the European consortium. The consortium will include AI and robotics companies and world experts in deep learning from across Europe.

Job requirements

Below is a list of skills that we would be happy to see in a candidate. Do not be intimidated by it. The must-haves are clearly tagged. For all the other points, consider applying even if you do not feel like you satisfy 100% of the requirements.

• First and foremost: scientific curiosity and passion! (must have)
• MSc in computer science, robotics, mechanical engineering, applied mathematics, physics, computer engineering or a related field (must have; fine if not achieved yet, but close to completion)
• Strong background in deep learning
• Familiarity with nonlinear dynamics and modeling mechanical or physical systems
• Strong programming skills in Python and PyTorch
• Experience with ROS and ROS2 is a plus
• Proficiency in written and spoken English (must have)
• Prior experience in machine learning for control, soft robotics, robot manipulation, or other forms of experimental robotics is a plus

TU Delft

TU Delft is a top international university combining science, engineering and design. It delivers world-class results in education, research and innovation to address challenges in the areas of energy, climate, mobility, health and digital society.

Faculty Mechanical Engineering

The Faculty of Mechanical Engineering is a dynamic and innovative environment with high-tech lab facilities and international reach.

Conditions of employment

Doctoral candidates will be offered a 4-year period of employment in principle, in the form of 2 employment contracts. An initial 1.5-year contract includes an official go/no-go progress assessment within 15 months, followed by an additional contract for the remaining 2.5 years assuming performance requirements are met.

As a PhD candidate you will be enrolled in the TU Delft Graduate School. The TU Delft Graduate School provides an inspiring research environment with an excellent team of supervisors, academic staff and a mentor. The Doctoral Education Programme is aimed at developing your transferable, discipline-related and research skills.

TU Delft offers a customisable compensation package, discounts on health insurance, and a monthly work costs contribution. Flexible work schedules can be arranged.

Will you need to relocate to the Netherlands for this job? TU Delft is committed to making your move as smooth as possible.

#EUfunded This is an EU funded project, named GRAIL, within program HE.