P1 Detect
Omen
Anomaly + RUL
Surfaces early failure signatures and remaining-useful-life signals from subsystem telemetry—so operators see drift before it becomes loss of mission.
Parallax
Ephemeris
Edge-AI predictive maintenance for spacecraft subsystems
Detect → Diagnose → Act
Mission-time autonomy is bottlenecked by ground-loop diagnostics. Ephemeris recaptures that budget with onboard inference at the edge.
Challenge
The problem
Mission-time autonomy for spacecraft is bottlenecked by ground-loop diagnostic cycles.
When a subsystem drifts out of nominal, operators often cycle through downlink, analysis, and uplink before acting. That latency consumes autonomy budget—and gives failures time to cascade across power, thermal, and avionics.
Approach
The solution
Ephemeris closes the loop on board, at the edge.
Telemetry flows into lightweight onboard inference. Each cycle produces three outputs: detect, diagnose, and act—before the anomaly becomes a mission-ending event.
Capabilities
Detect → Diagnose → Act
Ephemeris runs a three-stage onboard pipeline every inference cycle.
P2 Diagnose
Vigil
Subsystem attribution
Identifies which subsystem is responsible for an anomaly, with confidence—narrowing the search space that ground teams would otherwise explore manually.
P3 Act
Ultio
Safety-gated actions
Recommends next steps via rules and learned policy—always gated by safety constraints so autonomy never outruns operator intent.
Get in touch
Contact Parallax
Request a briefing or discuss collaboration on Ephemeris edge inference for spacecraft missions.
Demo form — submissions are stored locally for preview until Formspree is configured.