CosmicAI Postdoctoral Fellow

Thursday, October 16, 2025 3:00PM
Saturday, January 10, 2026 4:00PM

(Cosmic

AI) invites applications for two Postdoctoral Fellow positions in Artificial Intelligence (AI) for scientific discovery. Each position offers an opportunity to contribute to the development of foundational AI methodologies and their application to data-intensive challenges in astronomy and physics. The Institute’s overarching goal is to create transparent, interpretable, and scientifically grounded AI systems that enable robust inference, accelerate discovery, and deepen our understanding of the Universe.

These positions focus on developing methods that are explainable, trustworthy, and aligned with scientific reasoning. The Fellows will join an interdisciplinary community advancing research in symbolic reasoning, explainable and interpretable AI, optimization, Bayesian inference, and experimental design, all in the context of complex scientific data.

These positions are designed for early-career researchers who wish to bridge the gap between AI methodology and scientific application – developing novel theoretical insights while contributing to practical tools that support next-generation astronomical surveys and simulations. Fellows will have the freedom to pursue independent lines of research while collaborating closely with statisticians, computer scientists, astronomers, and physicists at the University of Texas at Austin and across the national Cosmic

AI network.

Applicants must specify for which position they wish to be considered. Each applicant may only apply for one position.

Advisors:
Dr. Alessandro Rinaldo and Dr. Arya Farahi

This position focuses on foundational and theoretical research in AI, emphasizing symbolic reasoning, interpretability, optimization theory, Bayesian inference, or experimental design. The Fellow will pursue new methodological advances that enhance the transparency, rigor, and reliability of AI systems used in scientific contexts. The ideal candidate will have a strong background in mathematical and statistical theory, algorithmic development, or the foundations of machine learning.

Example research areas include:

• Bayesian inference and uncertainty quantification
• Symbolic and hybrid symbolic and statistical reasoning
• Explainable and interpretable AI theory and methodology
• Experimental design and information-theoretic approaches to data efficiency
• Optimization methods for large-scale, non-convex, or domain-specific problems

This position focuses on the development and application of advanced AI models for complex scientific data, particularly in astronomy and physics-informed learning. The Fellow will contribute to creating and deploying interpretable, generative, and physics-constrained models that enable trustworthy scientific inference and discovery. The ideal candidate will have strong computational and modeling skills and an interest in interdisciplinary collaboration.

Example research areas include:

• Causal discovery in scientific data
• Explainable and interpretable deep learning
• Bayesian modeling and simulation-based inference
• Physics-informed neural networks (PINNs) and generative models

Fellows are expected to spend approximately 50% of their time on Institute-related collaborative research and 50% on independent research that aligns with the Institute’s goals. Ideally, there will be substantial synergy between the two. Appointments are for two years, with the possibility of renewal for a third year based on performance and funding.

Fellows will be based at the University of Texas at Austin and will work closely with interdisciplinary teams across the Cosmic

AI Institute. Each Fellow will have access to state-of-the-art GPU and CPU resources at the Texas Advanced Computing Center (TACC).

Ph.D. in Statistics, Computer Science, AI, Machine Learning, Applied Mathematics, Astronomy, Physics, or a related field by September 1, 2026.

Demonstrated expertise in one or more of the following:
Symbolic Reasoning, Explainable AI, Optimization, Bayesian…