Research, competitions, and applied work spanning reinforcement learning, intelligent tutoring, and adaptive VR cybersecurity training.
A reinforcement learning–driven VR training system that personalizes cybersecurity education. Uses PPO and AIRL to dynamically adjust challenge difficulty based on player performance. Developed as an NSF REU project at the University of Missouri–Columbia; findings presented at the NSF REU showcase.
An intelligent tutoring framework that enforces instructor-regulated learning through a three-layer gatekeeper access-control system (role-based, attribute-based, cognitive state-based) built on a knowledge graph of learner state and curricular content. Led UI development and co-authored the accompanying paper (ICHMS 2026), benchmarking SAGE at 90% curriculum access-control security vs. 30% for prompt-based tutors, with comparable helpfulness. Preparing for pilot deployment in Mizzou CS courses, Fall 2026.
An interactive 3D Mars map enabling users to plot waypoints on a rendered Martian surface. Implemented A* pathfinding with geodesic-based heuristics to compute optimal routes across spherical terrain, deployed for a live demo during the competition.