PhD Position on Agentic and System-1/System-2 Architectures for Robot Autonomy (AI Robotics)

Location: Turin, Italy; research conducted in collaboration with the MIRA Laboratory at AI4I and the Vandal Lab at Politecnico di Torino, including validation in laboratory settings on real robotic platforms. Working hours include CET for the application deadline.

Company

The Italian Institute of Artificial Intelligence for Industry (AI4I) conducts application-oriented artificial intelligence research focused on industrial innovation, technology transfer, and Industry 4.0.

What you will do

• Review literature and benchmark agentic AI, VLA models, world models, and System-1/System-2 architectures for embodied AI.
• Develop architectures in which multimodal reasoning models orchestrate VLA policies, learned skills, and model-based controllers.
• Study representations exchanged between deliberative and reactive levels, invocation strategies, time-scale separation, and safe low-level execution.
• Develop verification, failure-recovery, continual-learning, targeted-data-collection, and deployment-driven data-flywheel mechanisms.
• Evaluate methods in simulation and on real robotic platforms, focusing on long-horizon task success, robustness, failure recovery, computational latency, and data efficiency.

Requirements

• Master’s degree or eligible equivalent qualification; Italian qualifications require a minimum final grade of 95/110, while foreign qualifications must provide access to PhD studies in the awarding country and meet the applicable minimum average grade.
• Accepted English language certificate at CEFR B2, unless an exemption in the applicable call applies.
• Strong background in machine learning, robotics, computer vision, or control.
• Solid programming skills in Python and experience with deep learning frameworks such as PyTorch or JAX.
• Independence, research motivation, and strong analytical skills.
• Required application materials include a CV, academic records and transcripts for Bachelor’s and Master’s studies, a publication list if applicable, and contact details for at least two referees.
• Candidates who have not yet obtained the required degree or English certificate may apply if they fulfil the requirements by the deadlines specified in the call.

Nice to have

• Knowledge of foundation models, Vision-Language-Action models, reinforcement learning, imitation learning, or world models.
• Experience with robotic systems, ROS or ROS 2, MuJoCo or Isaac Sim, optimal control, or model predictive control.

Culture & Benefits

• Research collaboration between AI4I and Politecnico di Torino, with access to AI4I’s high-performance computing infrastructure and state-of-the-art laboratories.
• Opportunities for research visits at international partner institutions, including Stanford University, MIT, and ETH Zurich.
• Opportunities to publish at leading robotics, machine learning, and artificial intelligence conferences and journals.
• Opportunity to join Ellis Unit Turin activities and apply to become an Ellis PhD Student under the supervisors’ guidance.
• Competitive salaries, performance-based incentives, industrial collaborations, and an international research environment.

Hiring process

• This announcement is a preliminary expression of interest and does not replace the formal Politecnico di Torino PhD application.
• Candidates must also apply through the official Apply@polito platform for the second session of the 42nd PhD cycle by 16 November 2026 at 14:00 CET.
• Reference letters are not required at the expression-of-interest stage; referee contact details must be provided.