Research Scientist - Generalist Agent for Game Environments

About the Role

KRAFTON is pioneering the path to players' dreams with bold imagination and breakthrough technology, creating unforgettable worlds for fans across the globe. Since 2021, KRAFTON has been focusing on deep learning research, expanding its game production capabilities and making aggressive investments, including recruiting core talent.

The AI Research Division collaborates with various internal and external departments to provide AI solutions for diverse problems and develops proprietary deep learning technologies through in-house research. The primary directions are:

• Developing deep learning technologies to assist game production, thereby increasing efficiency and universality in game creation.
• Developing Multi-modal LLM-based Foundation Models that can play games alongside users.

R&D Focus

We research and develop Generalist Agents for Game Environments that can flexibly adapt to various goals and contexts. We aim to create agents capable of real-time decision-making and interaction based on an understanding of complex game environments. These agents should be able to comprehend diverse inputs such as text, visual information, game states, and logs, and generate appropriate actions and interactions for the situation. They should also operate flexibly according to changing environments and objectives, not confined to fixed roles or single play patterns. This involves handling the entire process, from agent architecture design to data construction, learning, evaluation, and inference optimization, to build systems with quality and scalability that can operate in real environments. Through this research and development, we aim to create new possibilities for the game-playing experience.

Culture Fit

Members of the AI Research Division interact and collaborate with team members from various fields through diverse projects, offering creative ideas for various problems. An atmosphere where opinions are freely expressed, regardless of age or rank, is encouraged. The team consists of individuals from diverse cultural backgrounds, and active support is provided to overcome language barriers, such as interpretation and translation, for smooth communication.

Responsibilities

• Research and development of Generalist Agents for Game Environments.
• Design and implementation of behavior and interaction generation agents.
• Performance improvement through data design, learning, and evaluation.
• Building and operating data pipelines.
• Real-time inference and serving optimization.
• Documentation and sharing of research results.

Required Experience

• Ability to join as a new Professional Research Personnel (전문연구요원) in 2026 (for those with a master's degree or higher, or who have completed a combined master's/doctoral program).
• Understanding of Agent systems and practical project experience.
• Proficiency in understanding deep learning papers and designing experiments.
• Project experience related to LLM, VLM, and multi-modal models.
• Ability to structure and solve problems.

Preferred Experience

• Experience publishing papers in major AI/ML conferences and journals.
• Research experience related to Agent, VLA, Decision Making, Human-AI Interaction.
• Experience with real-time inference optimization, serving, and service application.
• Experience with data-driven evaluation and improvement.