Location
Paty Vu
UCSF
521 Parnassus Avenue
San Francisco, CA 94158
United States
Job Posted Date
Opportunities
The University of California, San Francisco is seeking a highly motivated Postdoctoral Fellow to join the EpochAI Lab in the Department of Anesthesia and Perioperative Care, working at the intersection of artificial intelligence, sepsis care, and health data science under the leadership of Dr. Romain Pirracchio. This position is part of an NIH-funded multi-site R01 multisite project (including UC Berkeley, UPMC and UPenn) aiming to develop and validate a real-time, AI-driven clinical decision support system (CDSS) to guide early treatment decisions in patients with suspected community-onset lung sepsis (COLS).
The successful candidate will help design and implement advanced machine learning and causal inference models, including conditional average treatment effect (CATE) estimation and simulation-based evaluation of clinical impact. The fellow will also contribute to prospective validation efforts, integration of AI tools into the UCSF EHR (via HIPAC), and cross-disciplinary collaboration with clinicians, statisticians, engineers, and implementation scientists.
Key Responsibilities
- Develop and implement AI/ML algorithms to estimate individualized treatment effects using large-scale EHR data
- Collaborate with UCSF HIPAC and clinical teams to deploy and test models in real-time “silent mode”
- Assist with development of an interpretable and EHR-embedded CDSS for early treatment in lung sepsis
- Lead analyses for external validation at collaborating institutions (UPMC, UPenn)
- Co-author publications, conference abstracts, and data science workflows
- Participate in stakeholder engagement and iterative UI design processes
Job Requirements
- PhD (or equivalent) in computer science, statistics, biomedical informatics, computational biology, applied mathematics, epidemiology, or related fields
- Deep understanding of causal inference, heterogeneous treatment effect modeling, or ML frameworks
- Proficiency in programming languages such as R and/or Python
- Experience working with large-scale EHR data or real-world clinical datasets
- Familiarity with TMLE, Super Learner, CATE estimation methods (e.g., T-learner, X-learner), Reinforcement Learning, or generative modeling is highly desirable
- Excellent communication and collaboration skills, including ability to work across disciplines
- Track record of peer-reviewed publications in relevant domains
How to Apply
Please send CV, research interests, and supporting letters via email to [email protected].
Location