GDC 2025 Session Viewer
Empower QA Teams With Multi-Modal AI (Presented by NVIDIA)
Logan Herche (Principal Engineer, NVIDIA)
Natasha Anisimova (Senior Solutions Architect, NVIDIA)
Location: Room 3009, West Hall
Date: Thursday, March 20
Time: 5:00 pm - 6:00 pm
Pass Type: All Access Pass, Core Pass, Summits Pass, Expo Pass, Audio Pass, Indie Games Summit Pass - Get your pass now!
Track: Production & Team Leadership
Format: Sponsored Session
Vault Recording: TBD
Audience Level: All
As games grow increasingly complex, they present significant challenges in QA and data parsing throughout development and post-release. This talk highlights two critical solutions within the development pipeline. First, we explore how Visual Language Models (VLMs) can efficiently detect bugs in expansive open-world games, where manual inspection becomes impractical. Next, we discuss how retrieval augmented generation (RAG) with large language models (LLMs) can summarize the vast amounts of data generated by these systems, enabling developers to quickly identify key insights and address the most urgent issues in their games.
Takeaway
Attendees will gain valuable insights into how AI-driven solutions are transforming QA and data analysis in game development. They will learn about the use of VLMs to efficiently detect bugs in large, complex open-world games and the innovative role of RAG and LLMs in summarizing and simplifying vast amounts of system-generated data. Attendees will leave with a deeper understanding of the challenges associated with scaling QA processes and actionable strategies for using AI to improve efficiency, accuracy, and decision-making in their games.
Intended Audience
This talk is for game developers, QA professionals, data analysts, and AI engineers looking to enhance QA processes and streamline data analysis in game development.
