Postdoctoral Research Associate, Foundation Models and Causal Inference Scientific Discovery

Postdoctoral Research Associate, Large Foundation Models and Causal Inference for Scientific Discovery

The University of Virginia School of Data Science and the Reasoning and Knowledge Discovery (RISE) Lab invite applications for a Postdoctoral Research Associate position at the intersection of large foundation models, causal inference, and scientific discovery. The successful candidate will pursue a bidirectional research agenda: investigating how foundation models, including large language models and multimodal models, can support causal discovery, causal inference, scientific hypothesis generation, and experimental design;

and developing causal approaches that improve the reasoning, robustness, interpretability, fairness, and scientific reliability of foundation models. Research may include the development of new algorithms, theoretical frameworks, benchmarks, datasets, agentic systems, and evaluation methods. Potential applications span science, health, education, and other interdisciplinary domains. The position offers substantial opportunities to shape original research directions, collaborate with researchers across disciplines, mentor graduate students, publish in leading venues, and develop an independent research profile.

The Postdoctoral Research Associate will report to Sheng Li, PhD, and work closely with members of the RISE Lab and interdisciplinary collaborators at the University of Virginia and partner institutions.

Key Responsibilities

• Lead independent and collaborative research projects involving foundation models, causal inference, causal discovery, causal machine learning, and AI-enabled scientific discovery.
• Formulate research questions, develop novel methods and algorithms, and design rigorous computational experiments.
• Investigate how foundation models can incorporate scientific and domain knowledge to generate, refine, and evaluate causal hypotheses.
• Develop causal methods for improving the reasoning, trustworthiness, interpretability, robustness, safety, and generalizability of foundation models.
• Develop benchmarks, datasets, evaluation protocols, and reproducible research software.
• Prepare high-quality manuscripts for peer-reviewed conferences and journals.
• Mentor graduate students and provide guidance on research design, technical implementation, scientific writing, and presentations.
• Participate actively in interdisciplinary collaborations with researchers in data science, computer science, statistics, health, education, and other scientific domains.
• Contribute to research proposals, project reports, open-source software, and other scholarly products, as appropriate.
• Maintain high standards for research integrity, reproducibility, responsible AI, and ethical use of data and computational models.

Minimum Qualifications

• Doctoral degree (PhD or equivalent) in Data Science, Computer Science, Machine Learning, Statistics, Electrical and Computer Engineering, Information Science, or a closely related quantitative field. All doctoral requirements must be completed at the time of hire.
• Strong publication record commensurate with experience, demonstrating original research contributions.
• Demonstrated research expertise in at least one of the following areas:
• Foundation models, large language models, multimodal learning, natural language processing, generative AI, or deep learning; or
• Causal inference, causal discovery, causal machine learning, graphical models, experimental design, or related statistical methodology.
• Experience designing and conducting computational research, analyzing results, and communicating research findings.
• Ability to lead research projects with appropriate faculty guidance while working effectively as part of a collaborative team.
• Strong written and oral communication skills.
• Commitment to rigorous, reproducible, and ethical research practices.

Preferred Qualifications

• A strong publication record commensurate with career stage, particularly in leading AI, machine learning, natural language processing, or data-mining venues such as NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL, KDD, or comparable selective conferences and journals.
• Demonstrated research contributions connecting foundation models with causal inference, causal discovery, or scientific reasoning.
• Experience with one or more of the following foundation-model topics:
• Pretraining, post-training, fine-tuning, parameter-efficient adaptation, alignment, or evaluation;
• Model reasoning, agentic workflows, tool use, or knowledge integration;
• Large language models, multimodal foundation models, or scientific foundation models;
• Trustworthiness, safety, fairness, interpretability, robustness, or out-of-distribution generalization.
• Experience with one or more causal research areas, such as causal discovery, treatment-effect estimation, counterfactual reasoning, causal representation learning, mediation analysis, transportability, invariant learning, or causal experimental design.
• Experience working with large-scale datasets, GPU computing, or…