Model-Based Autonomous Land Rover
Led a five-person team building a model-based rover platform for GPS, odometry, sensor fusion, and autonomous waypoint navigation.
Led a five-person team building a model-based rover platform for GPS, odometry, sensor fusion, and autonomous waypoint navigation.
Developed and evaluated inference-time activation steering for pretrained vision-language-action models to guide safer robotic behavior without retraining.
Led motion-planning and electronics integration for an autonomous vehicle built with perception and hardware teams.
Designed a simulated underwater autonomy stack combining mapping, planning, control, and runtime safety constraints for cave exploration.
Built LEGO EV3 robotics projects spanning sensing, control, navigation, perception, and path planning.
Implemented and evaluated a DQN-based reinforcement-learning controller for CartPole stability.
Explored a learned-indexing approach for DuckDB inspired by ALEX, including offline rebuild behavior for analytical workloads.