Lead of Physical AI Agent, Research Scientist

Krafton

| Seoul, South Korea (On Site) | Full Time | 6 months ago
Apply Now

Job Summary

KRAFTON AI seeks a Research Scientist to lead the development of next-generation humanoid robot foundation models and embodied AI systems. The role involves designing and developing cutting-edge AI algorithms, optimizing agent behavior for physical environments using techniques like RL and imitation learning, building scalable training pipelines for large foundation models, and creating learning environments for real-world adaptation. The successful candidate will deploy models in simulation and on real robotic platforms, explore virtual-physical world interactions, and collaborate with cross-functional teams. Responsibilities include developing learning-based algorithms for manipulation, locomotion, and navigation, building scalable training and inference pipelines, and optimizing AI models for both simulation and real-world applications.

Must Have

  • Ph.D. in AI/related field or equiv. experience
  • Bilingual (English & Korean)
  • Strong engineering background (PyTorch, JAX, TensorFlow)
  • Proficiency in Python or C++
  • Experience with RL and imitation learning
  • Experience with large-scale AI/ML systems

Good to Have

  • Publications in top-tier conferences
  • Experience with IsaacSim, Mujoco, ROS
  • Strong data analysis skills
  • Experience with LLMs, Vision-Language Models
  • Knowledge of robot kinematics, dynamics
  • Experience with robot hardware

Job Description

 
크래프톤은 차별화된 크리에이티브를 가진 글로벌 제작 스튜디오들과 고유한 즐거움이 있는 게임을 발굴하여 전 세계에 퍼블리싱합니다.
우리는 지속적인 도전과 새로운 기술로 게임의 즐거움을 혁신하고, 더 많은 팬들에게 사랑받을 수 있도록 플랫폼과 서비스를 확장합니다.
우리는 열정과 의지를 지닌 동료들과 함께 성장하며, 글로벌 시장에서 인정받는 게임의 명가로 나아가고 있습니다.
 
 
KRAFTON is dedicated to identifying and publishing games worldwide that offer a uniquely enjoyable experience, crafted by our globally recognized production studios known for their distinctive creativity.
We continuously push the boundaries of gaming enjoyment by embracing new challenges and technologies, expanding our platforms and services to captivate an even larger fan base.
In collaboration with our passionate and driven team, we strive to establish ourselves as a renowned creator of masterpieces in the global gaming market.
 

About the job

  • KRAFTON AI is looking for an exceptional Research Scientist to help build the next-generation humanoid robot foundation models and embodied AI systems. Our mission is to develop general-purpose embodied agents that can learn, adapt, and master complex skills across both virtual and real-world environments.
  • You’ll be joining a highly collaborative and forward-thinking research team, known for impactful work in multimodal foundation models, large-scale robot learning, game AI, and physical simulation. In this role, you’ll have the opportunity to drive ambitious new research projects and contribute meaningfully to our long-term vision and product roadmaps.
  • KRAFTON Physical AI Team combines KRAFTON’s AI expertise with technologies that understand physical environments to research and develop Embodied AI - AI that can interact within the real world. Leveraging our experience with Co-Playable Character (CPC) and Reinforcement Learning (RL) technologies built through games, we create AI agents that deliver new kinds of user experiences. Ultimately, our goal is to expand KRAFTON’s AI research into the physical world and explore how AI can naturally collaborate and interact with humans.

Main Responsibilities 

  • Design and develop cutting-edge AI algorithms and models for general-purpose humanoid robots and embodied agents.
  • Research and optimize behavior and interaction models of AI agents tailored for physical environments, using techniques such as Imitation Learning, Reinforcement Learning (RL), Human-Robot Interaction (HRI), and Sim-to-Real Transfer.
  • Build scalable training and inference pipelines for large foundation models customized for embodied intelligence.
  • Develop learning environments that enable AI agents to adapt and respond autonomously in real-world spaces, leveraging KRAFTON’s AI technologies and in-game interaction experience.
  • Optimize and deploy AI models in both high-fidelity simulation environments and on real-world robotic platforms.
  • Explore new modes of interaction that bridge the virtual and physical worlds, applying research outcomes to real products and services.
  • Collaborate closely with cross-functional research and engineering teams across KRAFTON AI to bring innovative Sim-to-Real Transfer research into practical applications.

Key Qualifications

  • Ph.D. in Artificial Intelligence, Computer Science, Electrical Engineering, or a related field; or equivalent hands-on research experience.
  • Bilingual in English & Korean
  • Strong collaborative mindset with leadership skills to align and motivate cross-functional teams.
  • Strong engineering background with proven skills in rapid prototyping and working with modern model training frameworks (e.g., PyTorch, JAX, TensorFlow). Proficiency in Python or C++ is required
  • Experience in developing learning-based algorithms for manipulation, locomotion, and navigation using techniques such as RL and imitation learning.
  • Prior experience in one or more of the following domains is highly valued:
    • Multimodal Foundation Models
    • Deep understanding and hands-on experience in one or more of the following areas: LLMs; Large vision-language models; Vision-Language Models and Vision-Language-Action models.
    • Proven ability to work with large-scale AI/ML systems and distributed computing infrastructure.
    • Robotics
    • Experience in key areas such as foundation and diffusion models; Agent-based Learning, Sim-to-Real Transfer, Human-Robot Interaction. and Interactive Imitation Learning.
    • Deep understanding of robot kinematics, dynamics, and sensor integration.
    • Ability to safely operate robot hardware, lab equipment, and tools.
    • Knowledge of control methods, including PID, model predictive control, and whole-body control.
    • Practical knowledge of robot hardware design, prototyping, and hands-on assembly.

Preferred Qualifications

  • Publication record in top-tier machine learning and robotics conferences such as NeurIPS, ICML, ICLR, ICRA, IROS, RSS, or CoRL.
  • Hands-on experience with simulation platforms like IsaacSim and Mujoco, as well as robot systems based on ROS, with a proven track record of deploying end-to-end learning models.
  • Strong skills in analyzing and utilizing large-scale, high-dimensional data, particularly in the context of training and evaluating deep learning models.
  • Ability to collaborate effectively with cross-functional teams, bridging the gap between AI research and interactive content, and leveraging end-to-end learning pipelines to enhance AI agent behavior and interaction.

Location

  • Seoul, South Korea (Yeoksam, Center Field) 

Employment Type

  • Full-time

 

 

아래 안내 사항을 확인해주세요!

  • 장애인 및 국가 유공자 등 취업 보호 대상자는 관계 법령에 따라 우대합니다.
  • 지원서 내용 중 허위사실이 있는 경우에는 합격이 취소될 수 있습니다.
  • 채용 전형 중 궁금하신 사항은 크래프톤 채용 FAQ 내에서 확인하실 수 있습니다.

 

BE BOLD, LEARN AND WIN! 크래프톤의 성장과 도전의 스토리를 알아가고 싶다면? 

Please check the information below

  • Those eligible for an affirmative action program such as individuals with disabilities or with distinguished service to the state will be given hiring preference according to relevant laws
  • False statements in your resume may lead to the withdrawal of the employment offer
  • Please contact career@krafton.com for other inquiries

BE BOLD, LEARN AND WIN! Would you like to explore KRAFTON’s growth and adventure stories?

13 Skills Required For This Role

Team Management Cross Functional Cpp Game Texts Prototyping Pytorch Deep Learning Reinforcement Learning Python Kinematics Algorithms Tensorflow Machine Learning