Postdoctoral Appointee - Federated Learning for Foundation Models

The Mathematics and Computer Science Division (MCS) at Argonne National Laboratory is seeking a Postdoctoral Appointee to conduct cutting-edge research in machine learning, with a focus on the efficient training and deployment of foundation models in a federated learning framework. The Postdoctoral Appointee will work on the development and optimization of federated learning techniques to enable the training of large-scale foundation models across distributed clients, addressing key challenges such as data heterogeneity and communication efficiency.

The appointee will contribute to the conceptual framework, design, and implementation of federated learning architectures, with a particular emphasis on improving model performance, scaling across distributed systems, and ensuring privacy and security in data handling.

Position Requirements

Required skills and qualifications:

• Ph.D. (completed within the past 0-5 years) in computer science, statistics, data science, applied mathematics, operational research, or a related field.
• Proficiency in coding with Python and experience in C, C++, or other comparable languages.
• Strong background in machine learning techniques and familiarity with ML frameworks such as PyTorch, Jax, or TensorFlow.
• Proven ability to collaborate effectively with scientists, divisions, and external institutions, including universities and national laboratories.
• Excellent oral and written communication skills for engaging with all levels of the organization.
• Ability to model Argonne's core values of impact, safety, respect, impact, and teamwork.

Preferred skills and qualifications:

• Experience with federated learning, particularly in the context of training or deploying foundation models.
• Expertise in managing large-scale training datasets using GPU-enabled computing.
• Familiarity with privacy-preserving machine learning techniques.
• Experience with distributed computing, scaling machine learning models, or handling heterogeneous datasets.
• Knowledge of continual learning frameworks and strategies.
• A strong foundation in statistical methods, optimization, or game theory is a plus.

Job Family

Job Profile

Worker Type

Time Type