Applied AI / Local LLM R&D

GenWorld + Ollama

A local LLM research track exploring NPC memory, relationship rules, and world interaction structure using Ollama-backed runtime experiments.

Ollama NPC Memory Quest Rules StreamingAssets Data

Role

Used the R&D project to test how local LLM rules could influence NPC memory, relationship state, and world feedback inside Unity.

Question

Could a local LLM become meaningful gameplay structure instead of a disconnected dialogue gimmick?

Outcome

Established explicit data structures for memory, relationships, and quest interaction that made the experiment easier to inspect and evolve.

Why It Matters

The project shows how I translate AI ideas into runtime rules and inspectable game structure, even in R&D mode.

GenWorld world
GenWorld NPC mood
R&D angle: the point was not to claim a finished product, but to validate how local LLM structure could hold up when tied to actual game rules.
Implementation signal: memory rules, relationship rules, quest hooks, and data layout mattered more than a flashy single prompt.
Roles: applied AI prototyping, local LLM experimentation, and gameplay-system R&D.
Korean version: open the original Korean case study.