We are taking one or two PhD students for Fall 2027.
We look for people who have done some research already, who can name a question about people, models, or the two together that they want to spend several years on, and who are comfortable with behavioral experiments or systems work. Computer science, cognitive science, HCI, and statistics backgrounds all fit.
Apply through the Purdue CS PhD admissions page by December 1, 2026, and name Ilia Sucholutsky and the AI Thought Partner Lab in the "future research interests" section of your statement of purpose. The GRE is not required. The department decides after the deadline, so we do not hold interviews before you apply.
The lab opened at Purdue in August 2026, and the first projects follow the five research stories. On people, we are building systems that learn how one person represents a problem from a handful of interactions, and that show what they are still unsure about, using similarity and triplet judgments as the measurement.
On working with AI, we want to know whether a person is more capable after working with a partner than before. That means pre-registered studies with held-out tasks after the partner is gone, rather than judging the partnership by the quality of its joint output.
On thinking together, we are asking when human-AI partnerships that help each person also help the group discover and correct its errors, and when they fail to compose. The tools are group experiments, agent-based simulations, and measures from the collective-intelligence literature.
Each of these is a place where a new student could start.
Purdue undergraduates and master's students can join a running study for a semester.
Email Ilia to be added to the lab meeting invite.
Reach out by email with a short note about your research, a CV, and one or two papers you would like to discuss. We do not always have funded positions, but we are happy to support visiting researchers and collaborations.