Join the Algorithms and Applications for Partial Differential Equations (PDEs) team to develop new algorithms to advance science and engineering applications. This work will help shape the future Quantum-Centric Supercomputer for PDEs by combining classical numerical computing, scientific AI, and emerging quantum methods to improve accuracy and generalization while pursuing scalability and lower computational complexity. Projects are matched to your expertise and research background. The research PDEs underpin scientific and industrial problems in computational fluid dynamics, weather forecasting, fusion, healthcare, and finance. High fidelity simulations can be costly, while learned models must generalize across geometries, conditions, and scales. Quantum computing creates additional algorithmic possibilities. We will assess their practical value through applicability to real problems and comparison with strong classical methods. You may work on one or more of the following directions:
- Develop neural operators, physics informed models, or foundation models that learn from simulation data and generalize to new problems.
- Develop and evaluate quantum, AI, and hybrid quantum classical approaches to solving PDEs, including their accuracy, scalability, and computational complexity.
- Build reliable benchmarks that compare accuracy, generalization, runtime, and computational cost against numerical and AI baselines.
- Develop scalable multi-GPU implementations for training models on large datasets and solving PDEs on complex geometries with billions of degrees of freedom.
Our projects can draw on real engineering challenges and partner data. As one example, with Dallara our team developed an AI model that predicts aerodynamic fields around complex race car geometries and enables rapid exploration of design choices. The exact internship topic and methods will be agreed with the selected candidate based on the team’s research priorities and the candidate’s strengths.
What you will gain
You will work closely with IBM Research scientists and engineers, with opportunities to collaborate with academic and industrial partners. Depending on the project, you may use modern GPU and cloud infrastructure and quantum development tools. We support strong research outcomes, including publications and patent submissions when appropriate. Internship, Master’s thesis, and semester project arrangements may be considered where compatible with the university program.
Minimum qualifications
- Applicants must be enrolled in an MSc or PhD program in computer science, applied mathematics, physics, engineering, machine learning, or a related field throughout the internship.
- Strong programming skills in Python and familiarity with Linux.
- Solid foundations in at least one of AI, numerical simulations and PDEs, or quantum computing.
- Ability to formulate research questions, test ideas carefully, and communicate results clearly in English.
Preferred qualifications
- Enrollment in a PhD program is preferred.
- Experience in any of the following is valuable; you do not need to cover every area:
- Publications in top tier AI and quantum algorithms conferences and applied mathematics journals.
- Discretization methods (FEM, FV, or spectral methods), PDE solvers (CG, GMRES, or multigrid), and conditioning or preconditioning analysis.
- Neural operators, physics informed learning, foundation models for scientific data, or PyTorch.
- Quantum algorithms, quantum information, Qiskit, or analysis of quantum resources and complexity.
- High performance computing, distributed systems, C++, CUDA, or rigorous software and benchmark development.
Location and application
The position is based at IBM Research Zurich Laboratory. We welcome applications from universities in Switzerland and abroad. The expected internship duration is between three and six months, subject to the final project and hiring arrangements. Please submit your current CV and, if available, links to research papers, code, or projects that show your work. You may also include a short video introducing your interests and experience; it is entirely optional and its absence will not disadvantage your application. Shortlisted candidates can expect one or more technical discussions, a coding interview, and a conversation about project fit.
Further reading
