Life-course foundation models
A life becomes a document: attributes are words, events are sentences. Transformers from 8M to 540M parameters, pretrained on 7.5 billion records, tested against 24 prediction targets.
Enamul Hassan
PhD Candidate in Computer Science, Stony Brook University
Assistant Professor (on study leave), Shahjalal University of Science and Technology
I build foundation models for human life trajectories. With Steven Skiena, I pretrain transformers on population-scale records — the full Dutch national registers, covering 23 million people, and since 2026 longitudinal health records — and test how much of a life they can and cannot predict. The data never leaves its secure environment, so the work happens on site, in the Netherlands and the United Kingdom.
Email CV Résumé Google Scholar ORCID GitHub LinkedIn
I am looking for a Summer 2027 research internship in foundation models for sequential data, health AI, or computational social science.
News
- Passed my Research Proficiency Exam and advanced to PhD candidacy.
- Started modelling patient trajectories from longitudinal electronic health records.
- First-author manuscript on life-course foundation models submitted; under review.
- Embeddings of Nation-Level Social Networks appeared at Complex Networks 2025 (Springer).
- Research stay in the United Kingdom through January 2026, continuing the Dutch registry work.
- Joined Steven Skiena's group; on site at Statistics Netherlands over the summer.
- Began the PhD at Stony Brook with a Chairman's Fellowship.
Research
Nation-scale network embeddings
A population network of 17 million people and 1.4 billion typed edges, quantised into cluster tokens that mean the same thing from one year to the next.
Health trajectories
The same framing applied to clinical data: diagnoses, laboratory results and medications, ordered in time, as a sequence a model can read.
Selected publications
Foundation Models for Life-Course Prediction on Dutch Social Registry Dataunder review
E. Hassan, T. Pial, F. Hafner, A. Macanovic, L. Sage, T. Emery, A. van de Rijt, S. Skiena
Manuscript, 2026
Embeddings of Nation-Level Social Networks
T. Pial, F. Hafner, D. Handzlik, E. Hassan, L. Sage, A. Macanovic, T. Emery, A. van de Rijt, S. Skiena
Complex Networks & Their Applications XIV, Springer, 2025
Reducing Cyber Risks in IoT Device Resource Management: A Probabilistic Banker's Algorithm Approach
E. Hassan, G. Hossain, P. Vajpayee
IEEE World AI IoT Congress (AIIoT), 2024
EmoNoBa: A Dataset for Analyzing Fine-Grained Emotions on Noisy Bangla Texts
K. I. Islam, T. Yuvraz, M. S. Islam, E. Hassan
AACL-IJCNLP, 2022
Teaching and background
Before the PhD I spent six years on the Computer Science and Engineering faculty at Shahjalal University of Science and Technology in Bangladesh — Lecturer from 2017, Assistant Professor from 2019, now on study leave. I taught 25 courses across 68 offerings, supervised 45 final-year thesis students, led three funded projects building Bangla NLP datasets, and coached the university's teams to two ACM-ICPC World Finals.
Teaching and service → Projects and systems →
Contact
ehassan@cs.stonybrook.edu
Department of Computer Science, Stony Brook University, Stony Brook, NY 11794