In this project, we propose different techniques for developing Artificial Intelligences for videogames using Deep Learning models, with a special focus on the usage of neural networks for the development phase of a videogame. To reach this point, we started the creation of a Deep Learning development pipeline, capable of integrating neural networks inside of the Unity game engine initially developed in Python. For the Python part of the project, we used mostly Keras and Tensorflow for the neural network training, with a special focus on compatibility and ease of use. As background for our work, we carried out an extensive research about the current approaches inside and outside of the game development industry. We studied known cases where videogames are being used in the deep learning industry as a tool for research and experimentation, with well known examples such as AlphaStar, capable of defeating professional players on the game Starcraft II, or OpenAI Five, who has recently defeated the world champions on the game Dota 2. We have also analyzed the status of the development of Artificial Intelligence inside the videogames industry, and the usage of Deep Learning as a part of the development process of videogames. We went through famous examples of usage of Artificial Intelligence as part of game development, especially pointing out games relationed with Deep Learning in some way. Moreover, we have analyzed the current problems and flaws that a Deep Learning driven development of a game’s artificial intelligence may have, and suggested guidelines for either application on a production environment or further experimentation. Finally, we integrated the neural networks created on an Android videogame called HardBall, developed by From The Bench Games, S.L.. We used the Artificial Intelligence created to emulate the behavior of a Non-Playable Character inside of the game, with the intention to substitute the previous Artificial Intelligence with a neural network capable of learning and playing the game.