Company

Roboblackbox is a two-person company in Seattle, founded in July 2026 by Mehmet Cenk Ayberkin and Yuval Tzairi. Between us: thirteen years on Microsoft Defender and eighteen years in security software, building systems that record events on machines we do not control and make those records independently checkable.

That discipline is the foundation of what we are building for robots.

Autonomous robots already produce a great deal of runtime data. The problem is not that the data does not exist. It is that the relevant history is often fragmented across robot manufacturers, fleet managers, software vendors, integrators, and operators.

When behavior changes or something goes wrong, those systems answer different questions and are controlled by different parties.

We built Roboblackbox so the operator can maintain a continuous, independent production record alongside those systems.

We record what the fleet publishes and derive findings from that record after the fact.

We do not steer, stop, or warn a robot. We do not decide fault, certify that a fleet is safe, score a fleet, or set an insurance rate.

Every finding traces back to the recorded inputs and the derivation that produced it. We distinguish what was observed from what was derived, and every report states what the record cannot establish.

We would rather show a limit than imply a certainty.

Roboblackbox is running end to end on industrial recorder hardware: capturing runtime data read-only, preserving a tamper-evident record, deriving production findings, and reconstructing bounded incident windows from that existing record.

We are now preparing the system for its first production design partners.

We are looking for two or three operators running autonomous fleets in warehouses, factories, or logistics sites for a quarter-long design partnership this fall, and for insurance and claims teams willing to review the resulting reports with us.

We start with autonomous robot fleets in warehouses, factories, and logistics sites.

The longer-term ambition is to build the independent behavioral record for autonomous machines in production, whether they roll, manipulate, or walk.