The Digital Twins Lab at Columbia Business School invites applications for Postdoctoral Research Fellows to work at the frontier of digital twins, AI agents, and simulation-based decision-making.
The Digital Twins Lab focuses on building rigorous computational frameworks for simulating human behavior using large language models and agentic systems, with applications to experimentation, policy testing, marketing, synthetic data generation, and organizational decision-making. A central goal of the lab is to ensure that digital twin technologies are scientifically grounded, empirically validated, and responsibly deployed.
Research Fellows will have substantial intellectual independence in defining their research agenda and are expected to collaborate closely with faculty affiliated with the lab, including Professors Olivier Toubia, George Gui, and Tianyi Peng, as well as other affiliated faculty at Columbia Business School and across Columbia University. Fellows will also engage with industry and academic partners through benchmark competitions, shared datasets, and collaborative studies.
The position is a one-year appointment, renewable by mutual agreement for up to two years.
Research Areas of Interest
We seek candidates with demonstrated creativity and research excellence. Examples of relevant research areas include, but are not limited to:
- Digital twins of individuals or populations, including persona modeling, preference learning, and behavioral simulation
- LLM-based simulation and agentic systems, including single- and multi-agent environments
- Fine-tuning, training, and adaptation of AI models for simulation, prediction, and decision-making tasks
- Evaluation and benchmarking of AI agents, including predictive accuracy, generalization, and robustness
- Human-in-the-loop experimentation, computational social science, and hybrid qualitative–quantitative methods
- Ethics, governance, and responsible use of simulation technologies for testing and experimentation
Strong interest in connecting methodological development with real-world studies or datasets is especially valued.
Qualifications
Applicants must:
- Hold a Ph.D. (or have defended by the start date) in Computer Science, Marketing, Operations Research, Statistics, Economics, Electrical Engineering, Data Science, or a closely related field
- Demonstrate a record of outstanding research achievement and promise
- Have strong programming skills (Python required; experience with ML/LLMs preferred)
- Be comfortable working across disciplinary boundaries (e.g., ML + social science, simulation + experimentation)
Fellows are expected to contribute to the broader mission of the Digital Twins Lab, including:
- Participating in lab meetings and collaborative projects
- Helping organize workshops, benchmark competitions, or public-facing research artifacts
- Mentoring students and interacting with academic and industry partners
- Coordinating the lab’s digital twin competition, where multiple teams of researchers will compete on predicting responses to various studies submitted by companies, using synthetic data
Application Procedure
Applicants should submit the following materials:
- Cover letter
- Research statement describing past work and proposed research directions related to digital twins, AI agents, or simulation
- Curriculum vitae
- Two letters of recommendation
Please email all materials with the subject line “Postdoc – Digital Twins Lab” to [email protected]
Applicants should submit the cover letter, research statement, and CV in one email. Letters of recommendation may be sent separately. Applications will be reviewed on a rolling basis.