Implementing Deep Q-Learning (DQN) from Scratch Using RLax JAX Haiku and Optax to Train a CartPole Reinforcement Learning Agent