Research Associate in Dexterous Manipulation and Robot Learning

Location: South Kensington Campus

Are you a robot learning researcher eager to push the boundaries of dexterous manipulation? Join DexLab led by Dr. Rolandos Potamias and Stefanos Zafeiriou, to lead transformative projects spanning robot foundation models and real-world dexterous manipulation.

DexLab at Imperial is seeking a highly motivated and talented Postdoctoral Research Associate who has demonstrated competence in conducting cutting-edge robot learning research. The position focuses on building dexterous, high-degree-of-freedom (high-DoF) manipulation policies and robot foundation models, such as vision-language-action models (VLAs), learned from human demonstrations, but not limited to this. The researcher will work on a state-of-the-art robotic platform comprising Sharpa Wave dexterous hands and UR7e arms, contributing both to methodological advances and to the deployment of learned policies on real robots. The role integrates knowledge from robot learning, computer vision, generative modelling and control.

What you would be doing:

The post will focus on research topics in probabilistic programming. The specific details of the project are flexible and will be determined based on the interests and strengths of the successful candidate.

You will conduct original research at the intersection of dexterous manipulation, robot learning and foundation models, exploring how large-scale human demonstration data (e.g., video, motion capture and teleoperation) can be turned into capable, general-purpose robot behaviour. You will develop novel algorithms and policies for high-DoF hands and validate them on real hardware. You will collaborate with a team of experts in robotics, computer vision and machine learning. We strive to publish in top-tier conferences such as CVPR, NeurIPS, ICLR, ECCV, ICCV, ICRA and RSS.

What we are looking for:

• A strong track record in top conferences and journals in robot learning, machine learning or computer vision.
• Demonstrated robot learning experience (e.g., imitation learning, learning from demonstrations, VLAs / robot foundation models, or reinforcement learning) is essential.
• Strong PyTorch programming skills and working knowledge of ROS.
• Hands-on experience deploying learned policies on real robots; experience with dexterous hands, teleoperation or motion capture for demonstration collection.
• Applicants must hold a PhD in computer science, robotics, engineering or equivalent.

You should have proven research record and publications in the relevant areas.

See job description for a full list of requirements.

What we can offer:

• Work in an exciting and rapidly developing area, with access to state-of-the-art hardware including Sharpa Wave hands and UR7e arms. This is a chance to grow and establish yourself as a leader in this field.
• Sector-leading salary and remuneration package (including 43 days off a year).
• Disseminating your research in major robotics, machine learning and computer vision conferences is supported and travel costs covered.

Further Information

Full-time, fixed-term post for 12 months.

*Candidates who have not yet been officially awarded their PhD will be appointed as Research Assistant within the salary range £45,399 - £48,876 per annum. 

In addition to completing the online application candidates should attach: 

• A full CV with a list of all publications
• A research statement indicating what you see are interesting research issues relating to the above post and why your expertise is relevant

Informal enquiries related to the position should be directed to:

Dr Rolandos Potamias; r.potamias19@imperial.ac.uk

For queries regarding the application process contact Jamie Perrins: j.perrins@imperial.ac.uk

Closing Date: 4^th November 2026 (midnight)

Advert information