Postdoctoral Appointee – Advancing AI and LLMs Scientific Discovery

We are seeking a highly motivated Postdoctoral Appointee with a strong background in artificial intelligence and machine learning (AI/ML), with particular emphasis on the development and application of Large Language Models (LLMs) for scientific use cases. This position focuses on advancing LLM capabilities to address complex challenges across a range of scientific domains.

As part of a multidisciplinary team, the candidate will work at the intersection of AI/ML, domain sciences, and high-performance computing. The role requires a strong foundation in LLMs and machine learning, along with the ability and motivation to develop expertise in large-scale model training and scaling on HPC systems, as well as in handling the unique characteristics of scientific data, including large-scale numerical datasets, complex simulations, and multimodal information.

This position provides the opportunity to work with some of the world’s most advanced computing resources, including flagship exascale supercomputers such as Polaris and Aurora, while collaborating with leading researchers across diverse scientific disciplines. We encourage candidates who are passionate about applying state-of-the-art AI technologies to address critical scientific challenges to apply.

• Recently completed Ph.D. (typically within the last 0–5 years, or soon-to-be-completed) in Computer Science, Applied Mathematics, or a closely related field
• Design and optimize multimodal LLMs to encode, fuse, and reason over heterogeneous scientific data from diverse modalities such as numerical tables, text, and images
• Conduct large-scale LLM training, including pretraining, fine-tuning, RL tuning, and domain-specific adaptation on HPC systems
• Design and implement fine-tuning and RL strategies to optimize LLM alignment, performance, and reliability for scientific applications
• Develop and deploy autonomous LLM agents capable of reasoning, planning, and decision-making to support complex scientific workflows
• Implement alignment, safety, and reliability frameworks to ensure LLM outputs are accurate, trustworthy, and robust in scientific contexts
• Evaluate and benchmark LLM reasoning, cognitive capabilities, and generalization to support robust analysis, interpretation, and decision-making
• Apply conformal prediction and uncertainty quantification techniques to generate reliable confidence estimates and risk assessments for LLM outputs
• Disseminate research findings through publications in peer-reviewed journals and conferences
• Ability to model Argonne's core values of impact, safety, respect, integrity and teamwork