R&D Associate - Atomic manipulation AI agents

R&D Associate Staff - Atomic manipulation with AI agents

The Center for Nanophase Materials Sciences (CNMS) is seeking a Research & Development Associate Staff Scientist to support research directed towards developing novel AI agents that can be implemented on atomic scale microscopy platforms such as scanning tunneling microscopy (STM) and scanning transmission electron microscopy (STEM) as well as synthesis platforms such as pulsed laser deposition (PLD) to synthesize and optimize materials.

The focus will be on developing and improving the reliability of agentic AI platforms that can perform atomic manipulation in STM or STEM but will also branch into thin film synthesis such as PLD.

As a Research Associate, you will contribute significantly to research in these areas, bridging simulations with modern AI agents (for example, using reinforcement learning) to inform policies to drive atomic manipulation in microscopy experiments leading to discovery and creation of novel states of matter. In addition to fundamental science discovery, the research will pursue development of data pipelines using automated workflows, creation of multi-modal databases and novel ML‑approaches that allow integration of different theory, simulation, and multi-modal experimental protocols.

The research is designed to provide opportunities for development of your experience and scientific vision. The position resides in the Data Nano Analytics Group, Theory & Computation Section, Center for Nanophase Materials Sciences (CNMS), Physical Sciences Directorate (PSD) at ORNL.

• Design, implement, and deploy advanced agent-based AI frameworks (e.g., reinforcement learning, hierarchical policies, multimodal agents) and apply them to scanning probe microscopy platforms—including STM and STEM—to autonomously manipulate atoms and construct designer lattices with targeted quantum or electronic properties.
• Extend AI‑driven control approaches to thin‑film synthesis platforms
  , such as pulsed laser deposition (PLD), by developing closed‑loop optimization strategies for growth conditions, defect engineering, and in‑situ diagnostics.
• Integrate simulations, theory, and experiment by developing workflows that combine atomistic modeling, surrogate models, and agent policies to guide experimental decision‑making for materials discovery.
• Develop robust automation pipelines that orchestrate data acquisition, analysis, experimental control, and model retraining, enabling reproducible, scalable, and high‑reliability AI‑driven experimentation.
• Report and publish scientific results in peer‑reviewed journals in a timely manner.
• Present results at international scientific conferences and meetings.
• Deliver ORNL’s mission by aligning behaviors, priorities, and interactions with our core values of Impact, Integrity, Teamwork, Safety, and Service. Promote equal opportunity by fostering a respectful workplace – in how we treat one another, work together, and measure success.

• A PhD in Physics, Materials Science, Chemistry, Computer Science, or closely related field.
• Sound understanding of advanced ML concepts and architectures and hands‑on experience with open‑source AI/ML packages (such as PyTorch, scikit‑learn, Tensor Flow, JAX, etc.).
• Two or more years of experience in applying machine learning methods for instrument control, such as on a microscope or on a nanomaterials synthesis platform resulting in publishable scientific results.
• Two or more years of experience with state‑of‑the‑art machine learning methods such as reinforcement learning and Bayesian optimization.
• Experience with operating microscopy platforms (scanning probe or electron microscopy) and/or nanomaterial synthesis platforms (such as physical vapor deposition or molecular beam epitaxy).

• Strong understanding of concepts in solid‑state physics, ferroelectrics and/or 2D materials.
• Experience with advanced AI/ML methodologies relevant to autonomous science, such as generative models, causal inference, symbolic regression, or model‑based RL for scientific reasoning and materials design.
• An excellent record of productive and…