RofehAI
Earth is minutes away. Sometimes it isn't reachable at all.
A message to and from a Mars-class mission can take the better part of an hour round-trip — and a solar event or an antenna misalignment can cut that link off completely. Long-duration spaceflight carries real risks, like SANS and immune dysregulation, that today's monitoring checks against a population average rather than how a specific person has actually been trending. By the time Earth learns something is wrong, the crew has usually already had to notice it, work out what it means, and describe it themselves.
What this is: a working reference implementation of RofehAI's proposed architecture — a personalized, trajectory-based health-risk monitoring platform for long-duration spaceflight crews, starting with SANS (spaceflight-associated vision risk) and immune dysregulation. Includes an LLM briefing agent (single-crew and fleet-wide questions, with persistent memory) and a near-real-time clinical alert relay demonstrating detection under a simulated communication delay.
Who this is for
- Investors and reviewers evaluating the concept and technical approach
- Potential institutional partners (e.g. TRISH) assessing feasibility
- Collaborators and advisors reviewing the architecture
What this is not
- Not a clinically validated product — runs on synthetic demo data only
- Not a medical device, not for real clinical or diagnostic use
- Not a finished commercial product — concept / reference stage
How to use it
- Select a crew member from the roster on the left to review their SANS and immune-dysregulation risk trajectory, confidence, and driving signals
- Review any suggested countermeasure — the system only ever suggests; a person decides and logs the action
- Ask the briefing agent a question about a crew member (e.g. "how is this person trending?") — it reasons across active domains and cites its sources; try a follow-up question too, it remembers the conversation
- Use "Fleet-wide triage" in the sidebar to ask about the whole crew at once (e.g. "who needs review today?")
- Use "Clinical alerts" to run a detection sweep, toggle the simulated comms link, and review/acknowledge any alerts — this demonstrates near-real-time clinical relay under a simulated communication delay