AAA Game-Generating AI Development

The goal is to build an AI pipeline that can autonomously create full-scale, Action/Adventure titles on par with modern AAA releases. Every game the system outputs must support seamless open-world exploration and present a rich, branching storyline, all packaged for native PC play.

I already have high-level design notes and several benchmark titles that define the quality bar; what’s missing is the underlying generative engine. The solution should cover content synthesis end-to-end—level layouts, quests, dialogue, asset placement, balancing, and final build export—while remaining modular enough to slot into either Unreal Engine or Unity workflows.

Key deliverables
• A trained model (or cooperative model suite) that ingests design parameters and produces a playable PC build.
• Tooling scripts or a lightweight editor extension so designers can tweak variables (world size, narrative depth, difficulty curves) and regenerate assets on demand.
• At least one proof-of-concept game world demonstrating a cohesive map, quest chain, and cinematic intro rendered inside the chosen engine.
• Documentation that explains model architecture, dataset preparation, and steps to reproduce builds on a fresh machine.

Acceptance criteria
A reviewer unfamiliar with the code should be able to clone the repo, follow the docs, run a single command, and launch a PC executable that delivers 30+ minutes of uninterrupted, bug-free gameplay featuring exploration, combat, and branching dialogue.

Please keep the codebase well-commented, leverage GPU acceleration where sensible, and aim for a clean separation between generation logic and engine-specific hooks so future platform expansion remains straightforward.