I build systems that have to survive contact with physics: controllers, simulators, robot-evaluation infrastructure, and the evidence needed to trust them.
I study aerospace engineering at Arizona State University and work across nonlinear control, physical AI, machine vision, state estimation, and scientific software.
| ATLAS Controls Lab | A reproducible controls and simulation-fidelity laboratory for nonlinear quadrotors: geometric control, MPC, ESKF estimation, CBF-QP safety, system identification, and exact experiment replay. |
| Inspect Robots Analysis · Verifier | Auditable physical-AI evaluation through comparability-gated statistics, hierarchical uncertainty, temporal multi-camera evidence, conservative abstention, and replayable provenance. |
| ComFree-MPC · IRIS Lab | Low-level MuJoCo execution, control allocation, actuator dynamics, and validation infrastructure for a three-drone contact-manipulation stack, integrated beneath team-owned CasADi/IPOPT MPC. |
My broader research question is simple:
When the model and the machine disagree, what evidence tells us why?
That question keeps pulling me toward instrumented experiments, system identification, failure semantics, held-out validation, and software that makes every claim replayable.
Python MuJoCo JAX / MJX CasADi / IPOPT MATLAB / Simulink LabVIEW
I also build smaller, more human things. Lovebird is one of them.