Machine Learning and Material Science Research ScientistNew

We are seeking a highly motivated AI & Materials Researcher to join our discovery efforts and sit at the intersection of computational physics and modern machine learning., * End-to-End Discovery: Leverage AI and computational tools to identify novel materials in silico and work with experimentalists to synthesize them in the lab, and identify and solve the key scientific challenges in this process.

• Deeply understand existing physical property prediction pipelines (e.g., DFT, MD) to identify bottlenecks and opportunities for acceleration.
• Design and train advanced machine learning models (e.g., Graph Neural Networks, Equivariant Neural Networks) to approximate expensive quantum mechanical calculations with high fidelity and orders-of-magnitude faster inference.
• Utilize Large Language Models (LLMs) and multi-modal agents to parse scientific literature, plan synthesis recipes, and make reasoning-based decisions on experimental parameters.
• Implement active learning strategies to guide the search campaigns through vast chemical spaces.

While deep understanding of functional materials and in-silico property prediction is essential, this role goes beyond traditional modeling. You will design the machine learning architectures that accelerate our simulations and also have the opportunity to build the intelligent agents that drive our physical laboratory., In order to set you up for success as a Research Scientist at Google DeepMind, we look for the following skills and experience:

• Ph.D. in Materials Science, Physics, Chemistry, Computer Science, or a related field.
• Computational Physics: Experience working with atomistic simulation tools (e.g., VASP, LAMMPS, Quantum ESPRESSO) and theory (DFT, Molecular Dynamics).
• Computational Material Science: Experience working with materials databases and tools (e.g. Materials Project, GNoME, Pymatgen).
• Machine Learning Engineering: Proficiency in Python and deep learning frameworks (PyTorch, JAX, or TensorFlow). Experience developing models for physical systems (GNNs, Transformers).
• Strong programming skills for workflow management, data analysis, and tool automation.
• Excellent teamwork and communication skills, with a desire to work in a fast-paced, interdisciplinary collaborative environment.

In addition, the following would be an advantage:

• A track record of bridging the gap between computational prediction and experimental discovery.
• Experience with LLM post-training or designing agentic workflows.
• Experience with high-throughput computational workflows and running simulations on HPC or cloud infrastructure.
• A track record of publishing at the intersection of AI and Science (e.g., NeurIPS AI4Science, Nature Computational Science, etc.).

Science is at the heart of everything we do at Google DeepMind. From the beginning, we took inspiration from science to build better algorithms, and now, we want to use our toolkit to accelerate scientific discovery. By bringing together specialists with backgrounds in machine learning, computer science, physics, chemistry, biology and more, we're optimistic that we can build new methods that will push the boundaries of what is possible and help solve the biggest problems facing humanity., Google DeepMind (GDM) is pursuing a ground-breaking research program in materials, aiming to accelerate the discovery of new functional materials by combining the predictive power of artificial intelligence (AI) and computational simulation with automated experimentation. The team is establishing experimental capacity to create a closed-loop, AI-driven discovery engine. Computational simulation is critical for grounding the AI and providing quick in silico feedback before materials are sent off to the lab for experimental validation.