Open biomedical datasets
We provide curated, de-identified clinical and imaging datasets so students can do real computational research without needing a traditional lab network.
Hack4Health Labs
Student research in computational medicine
Hack4Health Labs is the research arm behind our hackathons. We democratize health research by giving students access to real biomedical datasets, mentorship, and a reproducible workflow — no lab network required.
Our approach
Real research, made accessible
Most students never get the chance to do hands-on biomedical research. Hack4Health Labs changes that. We structure each competition like a research program: a focused clinical question, curated datasets, mentorship, and a deliverable that can be reviewed and reproduced.
The goal isn't just a winning score — it's teaching students how rigorous, honest health research actually works.
What we provide
The building blocks of student research
Mentorship
Participants are paired with mentors who guide them through framing a question, building models, and writing up results.
Reproducible by design
Every project is submitted as a runnable notebook plus a written report, so results can be re-run and reviewed end to end.
Responsible methods
We emphasize data handling, model interpretability, and honest evaluation over leaderboard-only thinking.
Research programs
Current and past research
Each season tackles a different clinical challenge, with its own datasets, methods, and student outcomes.
Active · 2026 Season
Byte 2 Beat
Cardiovascular disease
A multi-month research hackathon challenging students to build machine-learning models that predict and interpret cardiovascular disease risk from de-identified clinical datasets.
Datasets
cardio_base · 70k recordsheart_processed · 918 recordsPTB-XL ECGCDC BRFSS
Methods
Risk predictionLogistic regressionRandom forestsModel interpretability
Completed · Oct 2025 – Jan 2026
AI 4 Alzheimer's
Neurodegenerative disease
An online research hackathon where students trained models for early detection and progression forecasting of Alzheimer's disease using MRI datasets such as the Augmented Alzheimer MRI Dataset.
Datasets
Augmented Alzheimer MRIStructural MRI scans
Methods
Image classificationDeep learningEarly detectionProgression forecasting
How it works
From question to reproducible result
Frame a question
Teams pick a clinical problem tied to the season's challenge and explore the provided datasets.
Build & experiment
Students engineer features and train models in notebooks, iterating with mentor feedback.
Submit & document
Each team submits a reproducible notebook and a report describing methodology and findings.