About the Team (Project)
KRAFTON AI Research Division provides AI solutions for various problems by collaborating with internal and external fields, and develops our own services through deep learning technology research. The direction is broadly twofold:
- Increase game production efficiency and universality through the development of deep learning technology that can help game production.
- Develop unique deep learning technology and apply it to various applications inside and outside the game.
The AI Companion Team applies LLM, SLM, and VLM in KRAFTON internal games to develop AI Companions that can interact with users, formulate strategies, and cooperate.
- The goal is for AI agents to become a team with players in a real-time game environment, breathe together, and provide an immersive play experience.
- The goal is to help stable adaptation by independently understanding game situations and providing proactive guidance to users.
- Proactively perform all processes from optimized model design to operation.
Our Mission
- Design innovative gameplay experiences using the latest AI technologies such as LLM, SLM, VLM, and Human-AI Interaction.
- Optimize the computational efficiency of various language/multimodal models that process text and visual input to respond to real-time interactions and large-scale concurrent access environments that occur in game environments.
- Collaborate with game designers and engineers to apply AI Companion to actual game projects and work together to improve player experience.
- Build and continuously improve feedback loops for model improvement through user data and game log analysis.
Required Experience
- Currently pursuing a master's or doctorate degree in Artificial Intelligence (AI), Computer Science, or related technical fields.
- Experience with one or more projects in LLM, SLM, VLM.
- Experience using Deep Learning frameworks such as PyTorch.
- No disqualification for overseas travel.
Preferred Experience
- Excellent research achievements such as first-author papers or awards published in renowned conferences (NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ACL, NAACL, etc.).
- Experience with language model serving using vLLM, llama-cpp, SGLang, etc.
- Experience using frameworks such as LLama Index, LangGraph.
- Experience with multimodal learning and inference.
- Excellent communication skills with other job groups (game planning, art, business, etc.).