Vienna Research

We are a team of clinicians, researchers, and engineers exploring how machine learning can be used to advance the mission of accessible mental health.

Already, millions of people are using ChatGPT and other AI assistants for mental health support. Early clinical trials show that LLMs can deliver meaningful reductions in PHQ-9 and GAD-7 scores. It is clear that generative AI will play a meaningful role in the future of mental health — and there are both significant risks and opportunities involved.

Our interest is in conducting novel psychotherapy research to uncover how and why psychotherapy works for whom. We believe that machine learning will allow us to understand granular details of the mechanisms of psychotherapeutic change, and thus, design new forms of mental health support that are more effective and scalable than the ones we have today.

About us

Eric & Suhaas

Hi! We were freshman year roommates at Harvard. We started Harvard Undergraduate Group Peer Therapy and spent much of college running these therapy groups. Suhaas is a Rhodes Scholar studying Machine Learning at Oxford. Eric started Harvard Tech for Social Good and previously worked at Google, Microsoft, and Watershed.

Prof. Bruce Wampold (Advisor)

World's foremost expert on therapy efficacy. Former President of the Society for Psychotherapy Research.

Prof. John Norcross (Advisor)

World's foremost expert on the therapist-client relationship. Former President of the APA Division of Psychotherapy.

Writing

Evaluating model behavior in crisis scenarios

April 9, 2026

We present an early framework for evaluating how models behave across crisis scenarios (suicidal ideation, self-harm, homicidal ideation, and psychosis).