Inspiration
Families managing rare and chronic conditions struggle most during flare periods. Caregivers must quickly decide what symptoms to track, how to adjust daily routines, when escalation is needed, and how to communicate changes to clinicians. Today, this process is manual, reactive, and stressful, and important health signals are often missed or poorly communicated.
Helios was inspired by the need to reduce this chaos and provide families with clear, conservative, and structured support during flare weeks.
What it does
Helios is an AI-assisted flare management system that transforms continuous patient signals into actionable guidance.
It provides:
- Personalized baseline profiling for each patient
- Continuous monitoring via AI calls, app logging, and wearable data
- Intelligent flare detection using personalized thresholds
- Daily flare checklists and routine adjustment guidance
- Clinician-ready summaries of what changed from baseline
The goal is to help families move from raw health data to clear next steps.
How we built it
Helios is built around a multi-stage pipeline:
Smart onboarding:
We collect baseline function profiles, daily routines, and baseline vs. flare care plans to establish each patient’s normal variability.
Continuous monitoring:
The system ingests data from AI agent phone check-ins, UI-based quick logging, and Apple Watch data to create a longitudinal signal stream.
Flare detection engine:
Daily inputs are continuously compared against the personalized baseline and classified into Green, Yellow, or Red states using conservative thresholds.
Guided response layer:
When meaningful changes are detected, Helios generates daily checklists, routine adjustments, and clinician summaries to support decision-making.
Challenges we ran into
Personal variability:
Every patient has a different “normal,” making population-based thresholds ineffective. We addressed this by focusing on individualized baseline modeling.
Making data actionable:
Many existing tools capture data but do not help families decide what to do. We focused heavily on structured outputs like checklists and summaries.
Multi-source integration:
Combining AI calls, UI logs, and wearable data into a unified pipeline required careful design to keep the system consistent and usable.
Accomplishments that we're proud of
- Built a personalized baseline system that adapts to each patient
- Designed a conservative flare detection framework with clear escalation levels
- Created a multi-channel logging pipeline (AI calls + UI + wearables)
- Generated clinician-ready summaries to improve communication
- Focused the product on caregiver usability during high-stress flare periods
What we learned
Through building Helios, we learned that:
- Care is increasingly happening at home
- Families need decision support, not just data collection
- Personalized baselines are essential in health monitoring
- Conservative escalation is critical in clinical-adjacent tools
- Multi-channel logging improves visibility into patient status
What's next for Helios
Next, we plan to:
- Expand wearable integrations
- Enhance multilingual support
- Improve flare detection accuracy with more longitudinal data
- Strengthen clinician-facing summaries and workflows
- Continue refining the user experience for families managing flare weeks
Log in or sign up for Devpost to join the conversation.