$ cat about.txtRobotics + VLA researcher working on sample efficiency in embodied learning — teaching robots to learn as fast as human brains.
- 🔬 Researching sample efficiency in embodied learning at Wilks H Honors (FAU HS) — published researcher
- 🤖 Building training-free robotic navigation on xArm6 + Ranger Mini 3, and converting a 2018 Gem e4 into a self-driving vehicle
- 🚀 Founder of AuthAI (healthcare prior-auth automation) · co-founder of The Ethics Lab
- 📚 Learning: statistical learning theory, sim-to-real transfer, computational neuroscience
- 🤝 Open to collaborating on: robot learning, VLA evaluation, anything at the intersection of neuroscience and embodied AI
- 📍 Boca Raton, FL — ET
- FLARE — Predicting vision-language-action policy failures from frozen activations. AUC 0.83–0.89, sub-millisecond overhead.
- Unlocking the Path — Robotic navigation that treats doors and elevators as permeable states — cheaper and more accurate.
- PhysMent — Accepted to the ICML 2026 AI4Math Workshop. A MuJoCo benchmark testing whether LLMs can reason about physics by actively experimenting with simulated environments.
- Aperion — Quantum wave simulation for portable stroke imaging, built in 48 hours at Qiskit Fall Fest 2026 and run on a real 156-qubit IBM Heron.
- Self-Driving Golf Cart — 2018 Gem e4 converted to full drive-by-wire on a unified ROS control stack.
- Who Do You Think You're Talking To — Longitudinal study showing AI identity disclosure significantly shapes relational communication.
