Reinforcement Learning Sim-To-Real Transfer of a Learned Policy From Isaac Lab to a Real SO-ARM101 System identification for stronger sim-to-real transfer on an open-source SO-ARM101. Writing My Own PPO Implementation and Making It Fast Getting my own PPO implementation correct, profiling it, and benchmarking it against the four RL libraries bundled with Isaac Lab. PPO on Atari Pong PPO implemented from scratch and trained on the ALE/Pong-v5 Gymnasium environment PPO on Lunar Lander PPO implemented from scratch and trained on the LunarLanderContinuous Gymnasium environment Deep Q-Network A DQN implementation trained on the Cart Pole Gymnasium environment Reinforcement Learning Implementations Q-Learning, SARSA, Actor-Critic with eligibility traces, and Deep Q-Learning across Gymnasium environments Robotics SO-ARM101 SmolVLA Finetuning Finetuning SmolVLA for real-world pick-and-place on the SO-ARM101 Visual Odometry and Visual SLAM on KITTI A stereo visual odometry pipeline extended with keyframing, loop closure, and pose-graph optimization RRT* with Differential-Drive Dynamics RRT* and kinodynamic RRT* implemented in Python notebooks PokerBot A Texas Hold'em playing robot combining YOLOv8 card vision, a DQN poker agent, and a suction-cup arm Hardware Animated Julia and Fatou Sets on an FPGA A VHDL VGA driver and fractal generator targeting an Intel DE10-Lite Underwater Robot Electrical Infrastructure Undergraduate capstone reviving an underwater robot's electrical systems and moving it onto ROS Other Advent of Code 2025 My Python solutions to the 2025 Advent of Code challenges
