About this hackathon
A 2-day hybrid hackathon with a €1,000 grand prize where teams (4 members maximum) build an AI/ML prototype that detects early, "weak" health risk signals by analyzing real-world, self-reported data - i.e., everyday inputs described by a person (not medical records). The goal is not just a model, but a product-ready prototype that could be piloted inside Hea.
Register below now (one application per team; if you don’t have a team yet, you can send an individual application via this form, and we will help you find one).
📆 Schedule
Day 1 (February 14th) - Kickoff & Build
• 09:00 - Introduction, task briefing, and Q&A
• 10:00 - Build starts
• 12:00 - Lunch
• 14:00 - Checkpoint #1
• 19:00 - Checkpoint #2
• 21:00 - End of day 1
Day 2 (February 15th) - Finalize & Demonstrate
• 09:00 - Continue building
• 12:00 - Lunch
• 14:00 - Checkpoint #3
• 16:00 - Submission deadline and team presentations (5 min per team)
• 19:00 - Сlosing
Awarding - February 18th
After considering the solutions, the committee of judges will select the top 5 teams and invite them to the final discussion and awards ceremony.
🤔 What you’ll build
Your solution should produce:
• Context-aware anomaly detection (vs the person’s own baseline, considering life events)
• Risk scoring + risk category (e.g., metabolic, psycho-emotional, and cardiovascular risk)
• A safe, empathetic follow-up question that gathers more context without sounding like a diagnosis
You’ll submit:
• Working prototype (Notebook or web-service) of your solution
• GitHub repo with source code
• Short team presentation (brief pitch format)
Rules & safety:
• No diagnosing or prescribing treatment (ethics requirement)
• Open source: at least 80% of code must be publicly available
🤝 Our partners
AltaIR CapitalHarbour.SpaceNebius
If you have any questions, feel free to reach out to us at n.bogachev@altair.vc
Terms & ConditionsThe Science Behind Early Health Signal Detection
Hosted on Luma