The Analytics Research Institute

Postdoctoral Researcher — Trustworthy and Responsible AI

The Analytics Research Institute (ARI) is seeking a postdoctoral researcher to develop and evaluate ethical, trustworthy, and safe uses of artificial intelligence in research analytics. ARI conducts research in the science of science and supports organizations in understanding research investments and their outcomes. This position will contribute to real-world projects, including work supporting the National Institutes of Health.

The researcher will investigate, and make recommendations on how to implement, how AI can support analysis of research grants, publications, and related information while maintaining rigor, transparency, and appropriate human oversight. The role combines hands-on development with research into reliability, bias, explainability, reproducibility, and the risks of relying on AI-generated outputs.

Working with ARI’s multidisciplinary team, the successful candidate will receive mentorship while developing increasing independence. Candidates should be self-starters and bring relevant research experience and an interest in applying it to practical problems. They are not expected to arrive with expertise in every method or data source.

Key responsibilities

• Design evaluations that compare AI approaches with established analytical methods and human review, using realistic tasks and clearly defined performance criteria.
• Investigate errors, bias, uncertainty, unsupported claims, and other failure modes; assess their consequences for intended users and applications.
• Develop and test safeguards, including source verification, human review, and criteria for when an output should be withheld or escalated.
• Create reproducible workflows and document data, methods, model configurations, supporting evidence, and limitations.
• Contribute to practical guidance for responsible AI use, including privacy, appropriate data access, transparency, and accountability.
• Communicate findings through technical documentation, demonstrations, presentations, and scholarly publications.

Required Qualifications

• A Ph.D. completed by the start date in computer science, data science, statistics, informatics, computational social science, or a related discipline.
• Research experience applying AI or machine learning, with practical experience in at least one relevant area, such as natural language processing, language models, embeddings, or retrieval.
• Proficiency in Python and experience developing documented, reproducible research code.
• A foundation in experimental design, model evaluation, and quantitative error analysis.
• Demonstrated interest in ethical, trustworthy, and safe AI, and an ability to critically assess model outputs and technical claims.
• Clear communication, collaborative working skills, and willingness to learn unfamiliar methods and subject matter.

Helpful, but not required

Experience with responsible-AI research, explainability, uncertainty estimation, human–AI interaction, knowledge graphs, or scientific and government data. Candidates may develop these skills through focused learning and project work.

Research and development opportunities

The position offers opportunities to shape reusable evaluation methods, publish findings, and gain experience translating AI research into tools and analyses that support real decisions.