The STELLAR project, funded by the Wellcome Trust and OpenAI, is seeking a Research Scientist to develop and validate Conceptor-based steering methods for large language models (LLMs), applied to realistic AI-powered patient simulations for mental health clinician training. Conceptors are a mathematical framework for precisely manipulating internal LLM representations, enabling continuous control over psychiatric symptom presentations that simple prompting cannot.

The position is at the NYU Stern School of Business in the Department of Technology, Operations, and Statistics. You will report to Joao Sedoc and will collaborate with other stakeholders as well as with an interdisciplinary team spanning ML, psychiatry, clinical speech, and public health.

We are looking for someone who wants to do careful, impactful work at the intersection of representation learning and clinical AI.

Responsibilities


+ Develop Conceptor-based steering methods for open-we...