For decades we have built and deployed reasoning machines for complex, high-stakes systems: AI that must infer and act, consider risk and reward, often in the physical world. Octet is the result of that experience, distilled into a primitive.
Octet is building the new primitive: proof of location: physical location you can sign, carry, and verify. Octet is our second venture-backed company; together we built and exited the first, focused on supply-chain optimization and alt-PNT. We've been working together for more than a decade, and part of our stack is built on positioning technology (geolocation without GPS, running entirely at the edge) that we first developed with the US Air Force; the technology is used today by special operations forces, navigating in GPS-denied terrain.
Backed by Variant, ex/ante, and hivemind, with angel investors in leadership at World, Coinbase, and quantitative hedge funds.
Began on the floor of the Chicago Mercantile Exchange, where he became the youngest senior trader in firm history. Has twice deployed first-of-a-kind AI, in financial markets and in mission-critical supply chains. Advised the UN and written on AI for the Financial Times.
MIT Professor, now Professor for Resource Aware Machine Learning at TU Munich. Awarded the €5M Alexander von Humboldt Professorship for Artificial Intelligence. His work in machine learning, optimization, and mathematics is widely cited. Suvrit writes production code; the core of Octet's engine is his.
Began his career in offshore logistics, from the Arctic to the Middle East, then a decade in edge technology where he led the roll-out of multi-billion-dollar supply chains and the first deployment of AI for supply chains across sub-Saharan Africa.
We're building the infrastructure to verify the real world — turning location into a new form of proof. Don't see your role here? Reach out and tell us why you want to work with Octet and what you'd like to build.