Agenda

This session explores the challenges of using AI programming tools in large-scale game projects and introduces a new differentiable AI solution. Attendees will learn how this solution addresses key pain points of traditional LLMs, including context limitations and hallucinations, enabling effective AI implementation in complex engineering environments. Practical case studies on automatic crash diagnosis and automated system development will be presented, offering actionable technical insights.

Takeaway

Attendees gain a new agent development paradigm for AI implementation in large-scale game projects, with practical case studies and actionable technical solutions.

Intended Audience

Target audience includes AI agent developers and game programmers using AI Copilot tools. Basic AI and game development knowledge is recommended.