Partners
Paragon Reflex partners with Tier-1 automotive suppliers, OEMs, robotics platforms, and eVTOL (electric vertical take-off and landing) programs. If you have to argue that your perception channel has an independent backup, we should be talking.
The reflex layer runs alongside your perception stack with its own model, its own training pipeline, and its own inference runtime. It integrates as a second detection input. Nothing in your stack changes to accept it.
Run our reflex layer alongside your existing perception model as an independent channel, without a new training pipeline and without changing your detector. The independence is from the perception software: both channels read the same sensor sweep.
Wheeled, legged, aerial. The resource budget is kilobytes of pattern state and a fixed execution cost, which is what makes a small automotive-grade part a plausible host. The port to a lockstep safety microcontroller has not been built yet.
The same layer runs on an NVIDIA Jetson board for small uncrewed aircraft. It has been measured on road data, not flight data. No airworthiness pathway is claimed.
What this is, and is not, today. Not a brake authority. Not in production in any vehicle. Today the channel monitors, arms, and captures events.
Tell us a one-line about your program (vehicle / robot / aircraft, target SOP, current safety bar, and what your team is stuck on). We’ll come back with a 30-minute intro and a mutual non-disclosure agreement (NDA) template if there’s a fit.
Email [email protected]