Tasking a satellite should be an optimization, not a shopping trip
Commercial Earth observation has never had more supply: dozens of constellations, optical and radar, sub-meter resolution, same-day delivery. But buying a collection still works like a catalog. A human browses vendors, compares brochure numbers, guesses at weather, and hopes the pass they picked was the right one.
Zerosat is the layer above the catalog. It predicts real orbital access from live ephemeris, folds in the actual weather forecast, prices options across every vendor, and selects the collection that best satisfies the requirement, with a written rationale for every decision. Not a marketplace: an optimizer.
Who is behind it
Zerosat is built by Khayre Ali in New York City. Khayre spent five years in finance before moving into computer science (M.S., Northeastern University), and builds applied-AI systems where decisions have to be explainable, priced, and audited. Zerosat is that discipline applied to satellite tasking.
Principles
- Real data or nothing: live orbits, live forecasts, live vendor pricing. Anything notional is labeled notional.
- Show the work: every recommendation carries its scoring math and the alternatives it beat.
- Humans approve spend: the optimizer proposes; a person approves every order.
- No tracking, no data resale, no dark patterns, on the product or on this site.