PhD: Safe Online Optimization for Dynamic Systems

Eindhoven University of Technology (TU/e); Nederland; On-site; EUR 36,000 - 45,000; Full time

Benefits offered by this job

ABP pension

Eight percent holiday allowance

Eight point three percent year-end-bon

Job summary

The Eindhoven University of Technology (TU/e) invites applications for a PhD candidate to develop safe Bayesian optimization for time-varying settings and distributed learning within the Control Systems group. The project focuses on theory with practical validation on drones and building climate control, under collaboration with Swiss partners.

You will contribute to algorithms, guarantees, and experiments, publish in leading venues, and participate in teaching tasks during four years.

Qualifications

• Master’s degree (or equivalent) in Control Engineering, Electrical Engineering, Mechanical Engineering, or Mathematics.
• If you are completing your MSc, indicate your expected graduation date in the cover letter.
• Research-oriented attitude with curiosity.
• Willingness to develop teaching skills and coach students.
• Programming knowledge (Python, Julia)?
• Fluent in spoken and written English (C1 level).

Responsibilities

• Develop safe Bayesian optimization algorithms for time-varying settings.
• Design distributed learning and optimization schemes with partial information and no central coordination.
• Study surrogate models and optimization to stay tractable in high dimensions.
• Derive convergence, regret and safety guarantees for non-stationary environments.
• Turn theory into algorithms validated on UAV reference tracking and building temperature control.

Skills

Python

Julia

English (C1)

Education

MSc in Control Engineering / Electrical Engineering / Mechanical Engineering / Mathematics

Job description

The Eindhoven University of Technology (TU/e) invites applications for a PhD candidate to develop safe Bayesian optimization for time-varying settings and distributed learning within the Control Systems group. The project focuses on theory with practical validation on drones and building climate control, under collaboration with Swiss partners.

You will contribute to algorithms, guarantees, and experiments, publish in leading venues, and participate in teaching tasks during four years.