Scientific Software Postdoc Autonomous Spectromicroscopy

Scientific Software Postdoc for Autonomous Spectromicroscopy

The Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory is seeking a Scientific Software Postdoc for Autonomous Spectromicroscopy to collaborate with an interdisciplinary team of scientists to perform original research and develop new automation capabilities for spectromicroscopy beamlines at the Advanced Light Source (ALS) by implementing machine learning-based sample screening and agentic AI-driven instrument control. The work will include developing and validating ML methods to identify high-value sample regions in real time;

developing an agentic AI component to enable automated instrument control; and evaluating system performance across a variety of experimental scenarios. The position will also entail communicating research results through peer-reviewed publications and scientific presentations, with the goal of enabling autonomous, intent-driven experimentation at x-ray spectromicroscopy beamlines.

The Advanced Light Source is a U.S. Department of Energy (DOE) Office of Science national scientific user facility that produces exceptionally bright soft and hard x-ray, ultraviolet, and infrared light. With a strong scientific reputation, expert staff, and advanced capabilities, the ALS attracts thousands of academic and industrial users each year in condensed matter and quantum materials, energy sciences, biosciences, earth and planetary sciences and more.

The ALS is one of five Berkeley Lab user facilities that serve 15,000 users annually. Co-located with the Molecular Foundry, NERSC supercomputing center, and Berkeley Lab's materials, chemical sciences, biosciences, and other divisions, it provides an ideal collaborative environment for innovative scientific discoveries.

The ALS is a global leader in soft x-ray science, and aims to maintain its leadership with ALS-U, a major project to upgrade the facility to a fourth-generation light source. This upgrade will position the facility among the brightest soft x-ray light sources in the world, offering capabilities that no other facility can provide. Following the ALS Upgrade, our beamlines will benefit from dramatically increased coherent soft x-ray flux, enabling microscale x-ray reflectometry to reach its full potential while already delivering impactful scientific results today.

You will:

• Investigate and validate machine learning or computational methods for real-time analysis of X-ray spectromicroscopy data at an ALS beamline.
• Create/extend an agentic component using modern AI techniques (tool and skill use, vision-language models, task planning) that interprets multimodal scientific input (images, text, etc.) to formulate instrument commands.
• Integrate this agentic component with real-time analysis results and beamline control into a unified system, and characterize its performance across a variety of experimental scenarios.
• Participate in experimental planning, commissioning, troubleshooting, and data-quality assessment associated with deploying new automation capabilities at beamlines.
• Apply appropriate human-in-the-loop, validation, and fail-safe approaches when developing AI-enabled experimental control workflows.
• Collaborate closely with scientists, engineers, and technical support staff at the ALS and with external academic collaborators.
• Contribute to existing open source scientific software and release relevant project outputs where appropriate.
• Present research findings at scientific meetings, conferences, and publish in peer-reviewed journals.

Required Qualifications:

• PhD degree in the Physical Sciences, Material Science, Applied Mathematics, Electrical Engineering, Computer Science, or related discipline.
• Experience and a strong interest in scientific software development or research software engineering.
• Knowledge of machine learning principles and practices.
• Familiarity with agentic AI systems and concepts.
• Ability to work collaboratively with a diverse team of scientists and engineers.
• Demonstrated record of scientific excellence through publications, talks, or software deliverables.
• Commitment to collaborative software development practices: version control and code review, unit testing, and continuous integration.
• Strong written and verbal communication skills, including the ability to write publications and present research findings.

Desired skills/knowledge:

• Proficiency in Python and the open source scientific Python software stack.
• Experience with hyperspectral or multidimensional…