Arturo Nunez

Arturo Nunez is the founder of Nereu, an AI-native game engine that lets people create playable worlds by describing characters, environments, and interactions. After nearly a decade at Unity and subsequent work at MongoDB, he is applying game-engine architecture and artificial intelligence to make development accessible without reducing creativity to one-shot generation.

From game engines to creative tools

Nunez began by developing Xbox 360 prototypes and web and mobile games at the Mexican studio Phyne. At Unity, he worked with developers in Latin America as a field engineer and technical evangelist, later serving as a developer advocate. His work included Unity’s integration with Visual Studio and the company’s Data-Oriented Technology Stack, including its entity-component system, job system, and Burst compiler.

After Unity, Nunez worked at MongoDB and helped a startup address version control for game assets. His public repositories document the technical interests behind Nereu: ECS_101 teaches entity-component architecture, ComputerGraphics explores real-time graphics, and AI-Study-Group collects artificial-intelligence learning resources.

Making game creation behave more like play

Nunez’s approach to AI-native game development replaces repetitive engine configuration with reusable systems that interpret a creator’s intent:

• Natural-language game creation: Nereu’s assistant, Bibi, can introduce a robot, configure keyboard movement and animation, add rain, and attach a following camera. Creators describe familiar gameplay concepts without manually connecting physics, rendering, or animation components.
• Asset tag system: Drawing on entity-component architecture, Nereu applies tags such as playable, animated, and drivable while reusable engine systems supply the corresponding behavior. Because behaviors are not tied to fixed object categories, even a building can become a controllable vehicle.
• Level-of-detail context management: Nunez adapts a graphics optimization principle to language-model context: nearby or relevant objects receive detailed descriptions, while distant scene elements are summarized or omitted. Vision models separately classify three-dimensional assets so creators can find objects by meaning.
• Game Context Protocol: His prizewinning prototype gives language models tools for creating geometry, materials, lighting, scenes, and effects; a request for snow can generate a configured particle system. His simplified 3D transform controls likewise prioritize common actions over exhaustive interface options.

Nereu runs in the browser and remains extensible through JavaScript. Nunez treats AI as support for experimentation and revision, while distinguishing interactive game engines from emerging world-model experiences that still face substantial real-time rendering and physics challenges.