Our world model targets general-purpose robot fleets, not single units. The moment you have more than one robot sharing a space, new problems show up that don't exist at the single-robot level: avoiding each other reliably, communicating what they're doing, and — ideally — splitting up and coordinating on shared tasks instead of just staying out of each other's way. You'd define how OMN4I thinks about multi-robot behavior from the ground up.
_ Research and build coordination mechanisms for multiple robots operating in shared space — starting with reliable collision and conflict avoidance as the floor, not the ceiling.
_ Work on inter-robot communication: what robots need to signal to each other, and how, for coordination to actually hold up outside simulation.
_ Push toward collaborative multi-robot task execution — cases where robots split, sequence, or jointly execute a task rather than just operating independently in the same space.
_ Connect this work to the core world model — how a shared or communicated model of the environment supports coordination, rather than treating it as a bolt-on layer.
_ Validate on our real robot fleet, not just simulation — coordination failures that only show up on hardware are exactly what you're here to catch.
_ Strong research background in multi-agent systems, multi-robot systems, distributed control, or multi-agent reinforcement learning — PhD or equivalent industry research experience.
_ Real experience with at least one concrete piece of this: collision/conflict avoidance, decentralized or distributed coordination, inter-agent communication protocols, or task allocation across multiple robots or agents.
_ Comfortable starting from "make sure they don't collide" and building up toward genuine collaboration — you don't need to have solved the hard version yet, but you should understand why the easy version isn't enough.
_ Can work close to real hardware and real physical constraints, not only in simulated multi-agent environments.
_ First-author publications on multi-robot systems, multi-agent RL, swarm robotics, or distributed coordination.
_ Experience with heterogeneous robot teams (different embodiments coordinating together), not just identical units.
_ Background connecting multi-agent coordination to learned world models or shared environment representations.
_ Track record of taking multi-robot work from simulation to real hardware.
