This project explores neuroevolution, combining the computational power of TensorFlow.js with the interactive visual capabilities of p5.js. Every five seconds, a new generation of agents is spawned, evolved from the highest-performing individuals of the previous cycle. Through selective reproduction and random mutation, these agents progressively refine their neural networks, adapting over time. Performance is evaluated by the height of the pole balanced atop a virtual hand—the higher the pole’s tip, the better the score—creating a dynamic, evolutionary pursuit of optimal control.