Developed a prioritized replay functionality for the ReplayBuffer in the WE-Autopilot/Red-Team repository, focusing on improving sample efficiency and training stability in reinforcement learning workflows. The solution introduced a new Sample class to store transitions with associated weights, enabling prioritized batch selection and dynamic updates during training. Persistence features were added to allow saving and loading of buffer states, supporting reproducibility and long-running experiments. The work leveraged Python and applied principles from algorithm design and data structures to enhance data efficiency. No bugs were reported during this period, reflecting a focused and robust implementation that directly addressed RL experimentation needs.