Mahzad Kalantari
Wizards of the CoastAbout
the speaker
Mahzad Kalantari is a machine learning and data science expert with a PhD in computer science and computer vision and over 20 years of experience in applied machine learning and data science. She works in the video game industry, where her focus is on integrating reinforcement learning into games to build intelligent systems, support game development, and enhance player experiences. She also applies machine learning techniques to automated game testing, helping teams improve quality, scalability, and production efficiency. She currently works at Wizards of the Coast, where she focuses on integrating reinforcement learning–based automated testing across video game studios.
Session Spotlight
From Theory to Practice: Reinforcement Learning in Video Games
Reinforcement learning (RL) is a powerful paradigm for training intelligent agents through interaction with their environment, yet its adoption in the video game industry remains limited. This presentation proposes a pragmatic, production-oriented approach to applying reinforcement learning to video games, illustrated through concrete examples in Unreal Engine.
The session begins with a high-level overview of the core principles of reinforcement learning (agent, environment, state, action, reward, and policy), providing insight into why RL is particularly well suited to interactive and dynamic systems such as games. It then focuses on the integration of reinforcement learning into Unreal Engine, covering architectural choices, data exchange with machine learning frameworks, and key practical challenges.
Finally, the presentation explores the use of reinforcement learning for automated video game testing, where RL-driven agents can explore environments, test game mechanics, and detect edge cases at scale. The goal is to demonstrate how reinforcement learning can evolve from research into a practical tool for game development and quality assurance.