Postdoc: Physics-Informed ML for Excited-State Chemistry

Lawrence Berkeley Labs; Berkeley (CA); Hybrid; USD 106,000 - 123,000; Full time

Job summary

Lawrence Berkeley National Laboratory is seeking a Postdoctoral Researcher in Scientific Machine Learning & Computational Chemistry to advance physics-informed, symmetry-aware ML models for nonadiabatic excited-state simulations. You will build equivariant graph neural networks and connect them to nonadiabatic dynamics workflows.

The role emphasizes developing robust research software, validating models on energy-relevant systems, and contributing to open-source tools while collaborating with

Qualifications

• PhD within the last 3 years in a technical field such as CS, physics, chemistry or applied math.
• Strong experience in machine learning for scientific data.
• Proficiency in Python with PyTorch or JAX.
• Knowledge of graph neural networks for molecular/physical systems.
• Ability to develop reliable software in Linux with version control and reproducible workflows.

Responsibilities

• Conduct fundamental research at the intersection of machine learning and computational chemistry.
• Develop equivariant graph neural networks for excited-state properties.
• Integrate models into open-source workflows and simulation interfaces.
• Publish results in peer-reviewed journals and present at conferences.
• Mentor students and collaborate across institutions.

Skills

Python programming

Machine learning

Graph neural networks

Linux environments

Independent research

Education

PhD in a technical field

Tools

PyTorch

JAX

DeepMD

NEXMD

i-PI

MixPI

NWChem

CP2K

Job description

Lawrence Berkeley National Laboratory is seeking a Postdoctoral Researcher in Scientific Machine Learning & Computational Chemistry to advance physics-informed, symmetry-aware ML models for nonadiabatic excited-state simulations. You will build equivariant graph neural networks and connect them to nonadiabatic dynamics workflows.

The role emphasizes developing robust research software, validating models on energy-relevant systems, and contributing to open-source tools while collaborating with