Devon Sandel-Fernandez, PhD, is a Clinical Psychologist and Postdoctoral Fellow, funded by the NIMH T32 National Research Service Award (NRSA) Integrated Mental Health Fellowship. Devon completed her doctoral training in clinical science at the University of California, Berkeley Department of Psychology, a graduate certificate in applied data science from the UC Berkeley School of Information, and her clinical psychology internship at the VA Puget Sound Health Care System in Seattle.

 

Devon conducts research predicting impulsive and risk behaviors as they occur in people’s daily lives. She has conducted numerous studies using ecological momentary assessment (EMA) and self-monitoring data from therapy to build person-specific models of symptom dynamics including self-harm, substance use, and suicide attempts. She has also contributed to the development of digital interventions for aggression and self-harm. Devon is now focused on the development and implementation of digital interventions which help people avoid engaging in behaviors they later regret. Devon is interested in the use of patient-generated data (e.g. EMA, passive sensing, social media use, AI chatbot conversations) to detect risk for behaviors like suicide. She also has experience training natural language understanding (NLU) and large language models (LLM) to gather diagnostic and risk information via AI chatbot interfaces.

 

Clinically, Devon provides care informed by cognitive behavioral therapy (CBT) and dialectical behavior therapy (DBT), and integrates digital tools to gather data, inform individualized treatment and monitor progress. She has worked with adolescents, families, and adults across university clinic, academic medical center, and VA hospital settings.

 

Beyond research and clinical work, Devon is passionate about teaching psychology and methods courses and has won awards for teaching at UC Berkeley, was a Graduate Remote Instruction Innovation Fellow, and received the Extraordinary Teaching in Extraordinary Times Award for course redesign for remote instruction.