MDplus

Programs · Datathon

The annual MDplus Datathon. Now in its 4th year.

A month-long national competition where medical students, residents, and physicians work in teams on real healthcare AI challenges. Past editions have published in peer-reviewed research and brought together hundreds of trainees from across the country and beyond.
4
annual editions since 2022
~300
participants in 2025
31
teams in the 4th datathon
1
peer-reviewed publication

Past editions

Four years of Datathons.

4th AnnualOct–Nov 2025

Empowering Patients Through AI

Largest and most internationally represented Datathon to date. Gold sponsors: Doximity, Inflo Health. Event partners: AMSA, ConductScience.

31 teams · ~300 participants

3rd AnnualOct–Nov 2024

Multiple competition tracks

First Datathon with multiple parallel tracks. Final pitch competition held November 18.

Multi-track format introduced

2nd AnnualOct–Nov 2023

Value-Based Care

28 teams using the MIMIC-IV dataset (via Hugging Face). Solidified the Datathon as the org's signature annual event.

28 teams · 71+ signups

InauguralJanuary 2022

National Inpatient Sample (NIS)

The first MDplus Datathon. Built and led by Lathan Liou (then Director of Data Science) and Katie Link (Director of AI). Established the model that's been published in peer-reviewed research.

200+ participants

The format

Three to four weeks. Six phases.

The same structure has run four years in a row, with refinements each cycle. Designed for clinicians who don't live in Jupyter notebooks.

Phase 1

Kickoff + team formation

Members opt in, mix into teams of 3–5, and pick from a slate of curated challenge tracks.

Phase 2

Workshops

Recorded technical sessions cover the dataset, tooling, and clinical context, designed for clinicians without deep ML backgrounds.

Phase 3

Build phase

Teams work for ~3 weeks on real clinical datasets. Mentors from sponsor organizations check in throughout.

Phase 4

First-round judging

Panel of physician-judges and industry reviewers scores submissions on a public 5-criteria rubric.

Phase 5

Finalist presentations

Top teams pitch live. Sponsors and the broader MDplus community watch and vote.

Phase 6

Award ceremony

Cash prizes for 1st, 2nd, 3rd plus superlative-category winners. Past winners have published, spun out startups, and joined the orgs that sponsored them.

Judging rubric

How submissions are scored.

Public 5-criteria rubric, max 38 points. We share it before the competition starts so teams know exactly what they're optimizing for.

  • Problem definition8 pts
  • Impact + outcomes6 pts
  • Practicality + implementation10 pts
  • Team capability8 pts
  • Presentation quality6 pts

Sponsors + partners

Backed by serious operators.

Sponsoring the Datathon is one of the cleanest ways to get in front of physician-trainees building in healthcare AI. Reach out if your team would like to be part of the 5th.

2025 Gold sponsors

  • Doximity
  • Inflo Health

Event partners

  • AMSA
  • ConductScience

Past sponsors

  • Scale AI
  • RStudio
  • Merck
  • Flatiron Health

Peer-reviewed

The Datathon model is a published medical-education paper.

“Leveraging Datathons to Teach AI in Undergraduate Medical Education.” Published in JMIR Medical Education in March 2025. The first peer-reviewed validation of the MDplus Datathon as a model for teaching applied AI to medical trainees.

See all MDplus publications

The 5th Datathon ships in fall 2026.

Join the community to get the announcement when applications open. The AI & Data sub-community is where it's organized.