Deep Learning for Optimization Researcher (PhD Internship)

Feyer is building AI for physical invention. We use differentiable simulators to invent new hardware setups in physics domains such as optics or quantum technology, and train models to explore these vast design spaces beyond the reach of conventional optimization and human intuition.

You’ll work on an ambitious research problem with significant compute and the opportunity to turn new ideas into real physical inventions.

This job is a paid 3–6 month PhD research internship, working 80–100% in Munich or Tübingen.

Qualifications

• Strong foundation in machine learning (e.g. you are working on a PhD in this field)
• Experience with and solid understanding of different optimization methods.
• Proficiency in Python and deep learning frameworks (e.g., PyTorch or JAX).
• You know how to use LLMs/agents to enhance your coding and research output and are able to communicate your own insights and decisions in simple, human terms.
• Collaborative mindset, willingness to work in an interdisciplinary team. Physics expertise is welcome but not required.