Team Idea
We’re building Fortifiers Physical Intelligence, a multimodal AI system for voice-directed bimanual manipulation. A user gives a natural spoken instruction, such as setting a dinner table with specific placement constraints, and the system converts that request into structured goals, object constraints, prerequisite steps, arm assignments, and executable actions. The agent then verifies the resulting scene and replans when necessary instead of blindly following a fixed sequence.
For the online Intel Bimanual VLA track, we’re combining multimodal scene understanding, task decomposition, causal planning, and adaptive dual-arm execution. We also plan to integrate Speechmatics so users can direct and correct the physical agent naturally through voice.
The goal is to build an agent that can understand what you want, reason about what must happen first, act with both arms, and verify that the requested state was actually achieved.
Submission
Submission draft in progress