Objective
Building toward a unified humanoid platform, but the proof is in metal first: systems I have fabricated, wired, debugged, and run under load before the full machine exists.
Anchor / proof
Completed hardware from co-ops and competition: machined assemblies, shop infrastructure, and mechanisms with measurable performance. Not renders; systems that shipped to a floor or a field.
Active R&D
Joint modules, limb architectures, and compute stack built on the same manufacturing discipline as the work above: FEA, gearbox simulation, bench-test plans, and tolerance stacks that must survive assembly and load.
Edge inference and motor bus orchestration: boot-handshake protocols, kernel bring-up, and low-latency sensing pipelines for closed-loop control.
Quasi-direct-drive cycloidal module: ~100 Nm spike torque, epitrochoidal profiles validated in FEA and prototype bench tests.
Serial 6-DOF limb design in progress: topology-optimized links, gearbox trade studies, and integration paths informed by competition-grade mechanisms already built and raced.
Competition mechanisms that set the reliability bar for end-effector and transmission design.
Sim-to-real policy work paired with HIL rigs from prior co-op work. Gait dynamics in simulation, failure modes caught on hardware before they reach a humanoid stack.
Physical test philosophy carried forward from Trexo co-op work.
Reach out for humanoid hardware development, structural optimization, or robotics chat.