Postdoctoral Appointee - Foundation Models Federated Learning

Core responsibilities include:

• Leading research on foundation models, including problem formulation, algorithmic development, and rigorous experimental evaluation.
• Advancing federated learning methods that enable distributed and privacy-aware training and adaptation of foundation models.
• Using modern AI tools to accelerate research productivity across ideation, coding, experimentation, analysis, and writing.
• Interpreting results critically and positioning contributions within the broader research literature.
• Publishing research outcomes and contributing to reusable research software when appropriate.

• PhD in computer science, applied mathematics, electrical engineering, statistics, or a closely related field, completed within the last 0–5 years is required.
• Demonstrated ability to conduct independent research, including problem formulation, methodological development, and publication in peer-reviewed venues.
• Strong background in machine learning, with research experience in deep learning, foundation models, or related areas.
• Solid programming ability in Python and experience with modern ML frameworks (e.g., PyTorch or equivalent), sufficient to support research and experimentation.
• Ability to effectively leverage modern AI tools to improve research productivity across the full research lifecycle.
• Strong written and oral communication skills, with the ability to publish research in peer-reviewed venues.
• Ability to model Argonne's core values of impact, safety, respect, integrity and teamwork.

• Prior research experience in federated learning, distributed learning, or privacy-preserving machine learning.
• Experience with large-scale model training or analysis of scaling behavior.
• Familiarity with challenges such as data heterogeneity, communication efficiency, or system constraints.
• Exposure to privacy, robustness, or security techniques (e.g., differential privacy, secure aggregation).
• Experience contributing to open-source research software.