Internship or Master’s Thesis

AI Solvers for Constrained Non-Convex Optimization

Ref. 2025_033

Motivation

Recent advances in AI-based solvers have demonstrated orders of magnitude acceleration in solving complex optimization problems compared to classical methods. To make these models practically relevant for engineering applications, the next frontier lies in developing architectures that can respect physical laws, handle continuous and integer constraints, and optimize non-linear objective functions. Such capabilities are essential to close the optimality gap while ensuring feasible solutions. This work will demonstrate the applicability of AI solvers to electric grid problems, including Optimal Power Flow (OPF), Expansion Planning, and Unit Commitment.

What You Will Do

• Explore AI-driven solvers for constrained non-convex optimization, including methods based on self-supervised primal-dual learning.
• Implement soft and hard constraints in neural architectures and evaluate the resulting models in terms of optimality gap and feasibility.
• Demonstrate and benchmark the AI solver on Optimal Power Flow and Unit Commitment problems against classical optimization solvers.
• Contribute high-quality, reproducible code to our open-source projects hosted under the Linux Foundation for Energy .
  

Required Qualifications

• For Master’s Thesis: Enrollment in a Master’s program in Computer Science, Electrical Engineering, Physics, Mathematics, or a related field.
• For Internship: MSc or PhD (preferred) in Computer Science, Electrical Engineering, Physics, Mathematics, or a related field.
• Hands-on experience with AI/ML model training, including understanding of optimization concepts, metrics, and regularization.
• Proficiency in Python and PyTorch, with solid software engineering skills (Linux, Git/GitHub, testing, reproducibility).
• Strong problem-solving mindset and ability to explore and synthesize ideas from literature, design experiments independently, and innovate new ML architectures.
  

Bonus Qualifications

• Foundational understanding of classical and data-driven optimization algorithms.
• Domain knowledge in electrical engineering or power system modeling.
  

Location, Timing, and Format

• Location: IBM Research Europe - Zurich, Switzerland
• Duration: Typically 6 months (Master’s Thesis) or 3-6 months (Internship)
• Supervision: Jointly supervised by researchers from the AI research teams, with access to IBM’s high-performance compute clusters.
  

Diversity

IBM is committed to diversity at the workplace. With us you will find an open, multicultural environment. Excellent flexible working arrangements enable all genders to strike the desired balance between their professional development and their personal lives.

How to apply

Please submit you application with a cover letter and curriculum vitae through the link below.

If you have any question related to this position, please contact Alban Puech, Alban.Puech2@ ibm.com .

Footer links