Aarhus University in Denmark is offering a three-year PhD fellowship in Physical AI and adaptive foundation models for robotics. The position focuses on developing intelligent robots that can interpret instructions, predict physical outcomes, and adapt to changing environments using advanced machine learning. The successful candidate will join the Adaptive & Agentic AI (A3) Lab within the Department of Electrical and Computer Engineering. Applications are open worldwide, including UK, EU, and international candidates, with competitive funding.
Important Dates / Deadline
- Application Deadline: November 1, 2026, at 23:59 CET
- Preferred Start Date: January 1, 2027
- Start Date Flexibility: January 2027 or later
- Duration: 3 years
- Study Mode: Full-time
Organization
University: Aarhus University
Graduate School: Graduate School of Technical Sciences
Department: Electrical and Computer Engineering
Research Laboratory: Adaptive & Agentic AI (A3) Lab
Main Supervisor: Associate Professor Behzad Bozorgtabar
Co-Supervisor: Professor Qi Zhang
Location / Study Destination
Aarhus, Denmark.
Who Can Apply
Country Eligibility: Open Worldwide, No Nationality Restriction Specified.
The PhD position welcomes applicants from the United Kingdom, European Union, and other countries worldwide. No nationality-specific restriction is indicated.
Applicants must meet the following qualifications:
- Hold a master’s degree equivalent to 120 ECTS in Computer Science, Electrical Engineering, Computer Engineering, Robotics, Machine Learning, or a related field.
- Complete the required master’s degree before enrollment.
- Demonstrate strong academic results and foundations in machine learning, linear algebra, probability, and optimization.
- Have strong programming skills in Python and PyTorch or a comparable deep-learning framework.
- Demonstrate substantial practical experience implementing, training, and evaluating deep-learning models.
- Provide evidence of research potential through a thesis, research project, code contribution, or publication.
Important: Applicants whose AI experience is limited to tutorials or using pretrained-model APIs will not meet the stated technical requirements.
Key Benefits / Funding
- Funding Type: Competitive PhD fellowship/scholarship.
- Position Duration: 3 years.
- Research Environment: Membership in the Adaptive & Agentic AI Lab.
- Academic Supervision: Guidance from researchers specializing in artificial intelligence and robotics.
- Research Resources: Opportunities to use public datasets, simulation environments, and available robotic or edge-computing platforms.
- Publication Opportunities: Research targeting leading international AI, computer vision, and robotics conferences.
The advertisement describes funding as competitive but does not provide a specific salary, stipend amount, or detailed funding package.
Main Requirements
The PhD focuses on building adaptive AI systems that connect perception, language, reasoning, and physical action.
Research will cover three interconnected areas:
1. Vision-Language-Action Models
Develop and evaluate models that connect visual observations and natural-language instructions to robotic actions.
Research topics include:
- Multimodal representations for robotic control.
- Learning from human demonstrations.
- Generalization to unfamiliar objects, tasks, and environments.
- Vision-language-action foundation models.
2. World Models and Planning
Develop predictive models that help robots anticipate the consequences of actions and plan effectively.
Potential research activities include:
- Learning physical dynamics from video and demonstrations.
- Model-based planning and decision-making.
- Learning through simulation and interaction.
- Transferring learned knowledge across different robot configurations.
3. Adaptation and Edge Intelligence
Develop efficient AI methods that maintain reliable robotic behavior under changing conditions.
Research topics include:
- Test-time adaptation and continual learning.
- Uncertainty-aware decision-making.
- Efficient inference and model updates.
- AI deployment under latency, memory, and energy constraints.
The precise research direction will be developed jointly with the successful candidate.
Research performance will be evaluated through task success, generalization, reliability, and computational efficiency.
The successful candidate is expected to conduct original research suitable for leading conferences such as NeurIPS, ICML, ICLR, CVPR, ICCV, CoRL, RSS, and ICRA.
How to Apply
Applications must be submitted through the official Aarhus University application system using the Apply button on the position advertisement.
Application steps:
- Confirm that you meet the master’s degree and technical eligibility requirements.
- Prepare evidence of practical deep-learning experience, including model implementation, training, and evaluation.
- Identify your individual technical contributions in previous research or development projects.
- Prepare the required application materials.
- Copy the project description provided in the advertisement and save it as a PDF.
- Upload the project description PDF as part of the application.
- Submit the online application by November 1, 2026, at 23:59 CET.
Documents Needed
The advertisement specifically requires evidence of the following:
- Project Description PDF: Copy the advertised project description and upload it as a PDF. This is a technical application requirement.
- Master’s Degree Qualification: Evidence of a completed or forthcoming relevant master’s degree equivalent to 120 ECTS.
- Research Experience: A substantial thesis, research project, code contribution, or publication demonstrating research potential.
- Technical Contribution Evidence: Clear identification of the applicant’s own contribution to previous research or deep-learning projects.
Important Notes
- Country Eligibility: Open to UK, EU, and international applicants worldwide. No nationality restriction is specified.
- Master’s Degree Requirement: A relevant 120 ECTS master’s degree or equivalent must be completed before enrollment.
- Strong Technical Experience Required: Applicants must demonstrate hands-on deep-learning model implementation, training, and evaluation.
- Research Evidence: Previous work must clearly show the applicant’s individual technical contribution.
- Project Description Requirement: Applicants must upload a PDF containing the advertised project description rather than preparing a new project proposal for this specific requirement.
- Research Flexibility: The final research focus will be agreed upon with the supervisory team.
- Funding Details: The position advertises competitive funding without specifying the monetary amount or detailed coverage.
- Publication Expectations: The project targets original research suitable for top-tier machine learning, computer vision, and robotics conferences.
- Application Deadline: Applications close at 23:59 Central European Time on November 1, 2026.