Research Scientist, Physics x AI

We are an AI Research Lab specializing in Physics x AI research. Our mission is to scale the capabilities of AI in the physical domain, encoding diverse scientific and engineering knowledge into neural networks to tackle complex, high‑impact industrial and scientific problems.

Location:

Princeton, NJ – you will work onsite in our office.

As a Physics x AI Research Scientist, you will explore and build the next generation of scalable, generalizable physics‑aware foundation models. You will focus on architectures (especially transformers) that move beyond narrow, hard‑coded models into robust solutions that generalize across geometries, conditions, and disciplines.

• Systems engineering and verification workflows (from design through manufacturing)
• AI‑native electronic design automation (EDA) and chip design
• Any discipline where physical or biological priors in AI can bring breakthrough scale and generalization

• Translate business needs into actionable project proposals, negotiate terms and budgets
• Steer delivery of complex, multinational strategic projects in cooperation with stakeholders and project teams
• Validate market potential of new AI innovations and develop business models for products, services, and solutions
• Drive ideas from ideation to productization with a customer mindset
• Own global Siemens cross‑company technology roadmaps and shape interfaces to other technologies
• Work with U.S. and global research groups to proactively shape the strategy of our AI research and pre‑development operations
• Shape the subsidies and grants portfolio for our AI research in the U.S.
• Communicate complex technology topics to non‑experts and executive decision makers
• Conduct hands‑on research and development of scalable, physics‑aware foundation models
• Model and encode knowledge from diverse scientific and engineering domains – including fluid dynamics, electromagnetics, electrical engineering, biology/protein folding (excluding quantum science)
• Document and publish research, and present proposals and results confidently to internal and external audiences
• Engage with Siemens domain experts, customers and partners to understand real‑world problems, collect requirements, and co‑create innovative solutions
• Translate cutting‑edge research into industrial impact, including proof‑of‑concept projects and identifying research gaps
• Identify, adopt, and apply evolving trends in machine learning, physics‑informed AI, and AI for science
• Collaborate with our interdisciplinary Princeton‑based lab and contribute to Siemens’ global Data Analytics & AI network

• Master’s or PhD in a STEM field (e.g., Computer Science, Physics, Applied Mathematics, Electrical Engineering, Mechanical Engineering, Biophysics, Bioinformatics, or similar)
• Established track record of publication or patent applications
• Proficiency in modern programming languages (Python, C++)
• Experience with AI libraries:
  PyTorch, JAX, NVIDIA NeMO, HF Transformers
• Proficiency in English, written and verbal
• Authorized to work in the United States without corporate sponsorship now or in the future (current Siemens employees on visa will be considered)
• Ability to work with controlled technology in accordance with U.S. Export Control Law; you may be required to provide citizenship status for compliance

• Proven ability to translate state‑of‑the‑art research into real‑world business impact
• Strong publication record (e.g., NeurIPS, ICML, IEEE, AAAI, L4DC) or equivalent industrial research achievements in industrial AI, physics‑informed AI, physical AI, or AI for science
• Experience training neural architectures capable of learning and/or encoding priors from physics, chemistry, biology, and/or engineering
• Experience with large‑scale, distributed AI/foundation model training
• Fundamental knowledge of software engineering principles and modern software development methodologies
• Ability to present research proposals and results to large, diverse technical and business audiences
• Familiarity with AI‑based surrogate modelling for CFD, AI‑native chip design, protein folding, or verification workflows
• Domain knowledge in…