• Supervised by Marzyeh Ghassemi.
• Research on foundation model evaluation, multimodal learning from longitudinal and wearable health data, and scalable AI methods for biomedical applications, with a focus on reliable and trustworthy AI, clinically grounded evaluation, and privacy and memorization risks in foundation models.
• I am happy to work with UROPs and other students, and I welcome interdisciplinary collaborations across computer science, law, medicine, and with different actors across the healthcare sector. Please add a description or high-level outline of your project idea to your inquiry.

• Supervised by Fredrik D. Johansson and Devdatt Dubashi.
• Presented several papers at conferences such as ICML, AISTATS, AAAI, and ML4H, with contributions published in two Alzheimer’s disease research journals, while leading multiple research initiatives in collaboration with academic and healthcare partners, resulting in eight peer-reviewed publications in top conference proceedings and journals.
• Supervised several master’s thesis projects, guiding students through academic research and industry collaborations.
• Served as a teaching assistant for multiple courses, including Introduction to Data Science, Research Methods, Causality and Causal Inference, and Design of AI Systems.
• Completed advanced coursework in Deep Learning, Software and Cloud Computing, Graphical Models, Bayesian Learning and Statistical Relational Learning, Artificial Intelligence and Machine Learning, and Causal Inference, developing a strong foundation in cutting-edge AI/ML techniques.

Thesis title:

• Time series analysis for the recognition of behavioral parameters in people with dementia

• Memorization in foundation models in health care
• Human-LLM Parity
• Predicting health outcomes from environmental and socioeconomic data

• Predictions under missing values in the context of TraumaBase
• User study to evaluate interpretable machine learning and missing values