Join us
We train scientists and clinicians to build clinical AI that understands time, place, and the limits of its own data.
CLAI members work with electronic health records from dozens of hospitals, run AI agents on hardware inside the institution, and publish in both clinical and computational venues. Postdoctoral fellows, PhD students, medical students, and undergraduates join the group, and each person works on a project that matters to patients and to the field.
What you could work on
Agents that read the clinical record
Multi-agent systems that read notes, critique each other's answers, and optimize their own prompts while running on local NVIDIA hardware, along with the study of how such systems fail.
How diseases follow one another
Temporal knowledge bases that map the order in which conditions appear across millions of patients, and models that reason over that order.
Environment, aging, and dementia
Linking extreme heat, air pollution, and other exposures to clinical records to study how the environment shapes health in aging adults and people living with dementia.
Precision cohorts and recovery
AI-defined cohorts across 58 hospitals that test treatment ideas in records, design better trials, and measure recovery after acute illness.
You will build on our open-source software, on data from Mass General Brigham and national research networks, and on close mentorship from the group’s director.
Postdoctoral fellows
We look for postdocs with backgrounds in computer science, statistics, bioinformatics, or machine learning. Experience with large language models and agentic systems, temporal and sequence modeling, causal inference, federated learning, knowledge graphs, or environmental exposure science fits our current work well. There are three routes into the group.
Apply directly. Email the following to hestiri at mgh dot harvard dot edu:
- A one-page research statement that describes the areas of machine learning you have worked on, how they connect to our research, and your preferred start date.
- Your CV with a list of publications.
- Links to two or three representative publications.
- Contact details for at least two references, such as former research supervisors.
Harvard Data Science Initiative. CLAI is affiliated with the Harvard Data Science Initiative, which selects several independent postdoctoral fellows each year for two to three years of research with Harvard faculty. Applications open each fall.
Eric and Wendy Schmidt Center at the Broad Institute. The Schmidt Center fellowship supports postdocs working where machine learning meets biology and medicine, with a mentor chosen from Broad-, MIT-, or Harvard-affiliated faculty. Hossein Estiri is an associate member of the Broad Institute, so fellows can choose CLAI. If you apply, name CLAI in your statement.
PhD students
We accept rotation and thesis students who are already admitted to a Harvard or MIT PhD program, including Bioinformatics and Integrative Genomics, Systems Biology, Biological and Biomedical Sciences, Harvard Integrated Life Sciences, and the Harvard-MIT Health Sciences and Technology program. Email your CV and the term in which you would like to rotate. Jonas Hügel completed a PhD co-advised by Hossein Estiri and is now a researcher at the Technical University of Denmark, and Sina Rashidian completed his PhD research and a postdoctoral fellowship with CLAI before moving to industry.
Medical students
Medical students on a research year, or completing a scholarly project, can lead a study within one of the projects above. Clinical training is an asset here, because many of our questions start at the bedside and end in the record. Alaleh Azhir completed her Harvard-MIT HST MD thesis research with CLAI and now continues that work as a cardiology fellow and CLAI investigator. Email your CV, your school and program, and the months you have available.
Undergraduates
Undergraduates with programming experience in Python or R can join for a summer or a semester and contribute to software, data pipelines, and analyses. Email your CV, a short description of something you have built, and your available dates.
How we review applications
We read every application, and we review postdoc applications on a rolling basis. Our capacity is limited, so we reply to candidates whose interests match a current project, and we invite shortlisted candidates to interview in person or by video.