Games have long served as proving grounds for AI. KRAFTON’s virtual worlds host millions of players whose behavior is unpredictable, social and constantly changing, and building intelligent systems for those environments has led the company to invest heavily in machine learning, reinforcement learning, large language models and multimodal AI. That work includes proprietary AI foundation models and Co-Playable Characters (CPCs) developed in collaboration with NVIDIA, AI teammates designed to understand players and respond to them in real time. Ludo Robotics builds on that foundation to take the same kind of intelligence into the physical world. Today, the company presented Ludi 0.2, the latest version of its robotic intelligence system, at the 2026 Bay Area Robotics Symposium (BARS) at UC Berkeley.
“For nearly two decades, KRAFTON has built games and virtual worlds where people and intelligent systems interact in real time,” said CH Kim, CEO of KRAFTON and Ludo Robotics. “Through Ludo Robotics, we’re building on that experience to explore how AI can better understand people and interact with them naturally in the physical world.”
Advancing Socially Intelligent Robotics Through Ludi 0.2
Ludo Robotics is developing general-purpose robotic intelligence systems designed to understand human intent, adapt to changing environments and collaborate naturally with people. Its approach focuses on a critical challenge in robotics: enabling intelligent systems to move beyond executing predefined commands to better understand context, interpret human behavior and respond appropriately as situations evolve.
Introduced in August 2026, Ludi 0.1 combined a vision-language model for reasoning and dialogue with specialized navigation and manipulation capabilities. The system was designed to interpret spoken instructions, maintain conversational context, ask clarifying questions and adapt to changing tasks within supported environments.
Building on that foundation, Ludi 0.2 advances several capabilities central to more natural human-robot interaction, including:
- Social Intelligence: Improving how the system interprets human intent, responds to feedback and accounts for social cues during interactions.
- Memory and Reasoning: Advancing the system’s ability to maintain context and reason through changing tasks and situations.
- Tool Coordination: Supporting more coordinated use of capabilities involved in navigation, manipulation and interaction.
- Gesture Interpretation: Improving the system’s ability to recognize and respond to human gestures as part of physical interaction.
- Physical Movement: Advancing smoother, more adaptable movement in supported environments.
Together, these capabilities represent continued progress toward robotic systems that can integrate communication, reasoning and physical action, allowing them to interact more naturally with people.
“Robots are becoming increasingly capable of performing physical tasks, but understanding and responding to people remains a major challenge,” said Kangwook Lee, Chief AI Officer of KRAFTON and CTO of Ludo Robotics. “Ludi 0.2 represents another step toward our goal of bringing conversation, reasoning and physical action together so robots can better understand human intent, respond to feedback and adapt as situations change.”
The company’s longer-term research roadmap includes Ludi 1.0, a next-generation foundation model currently in development that aims to integrate perception, dialogue and physical action within a unified system. The goal is to enable robots to maintain a shared understanding of tasks with people while communicating and acting in real time.
By bringing these capabilities together, Ludo Robotics aims to advance robotic intelligence that can support more natural collaboration between humans and machines.
Establishing a Silicon Valley Hub for AI and Robotics Research
Ludo Robotics is headquartered in Palo Alto, with an additional R&D lab in Seoul, Korea. The location places the team close to the Bay Area’s deep research talent, leading academic institutions and broader technology ecosystem.
The company’s scientific leadership brings together expertise across AI, robotics and human-computer interaction. Chief Technology Officer Kangwook Lee, who also serves as KRAFTON’s Chief AI Officer, is a former tenured associate professor at the University of Wisconsin–Madison whose research spans machine learning, large language models, reinforcement learning and multimodal AI. Chief Scientist Rob Nowak is an IEEE Fellow and UW–Madison professor specializing in machine learning, embodied AI and robot learning, and Human-Robot Interaction Lead Yea-Seul Kim is a former UW–Madison assistant professor of Computer Sciences focused on human-computer interaction, information visualization and data science.
Together, the Palo Alto and Seoul locations will support Ludo Robotics’ continued research into socially intelligent robots, with the long-term ambition of developing robotic systems that can collaborate naturally with people in homes, workplaces and other real-world environments.
Building the Research Foundation for socially intelligent robots
Real-world human interactions involve ambiguity, changing preferences, non-verbal communication and unpredictable behavior, which makes them difficult to train robots for. To address this, Ludo Robotics combines simulated environments, AI-generated social interactions, and real-world human behavior data. The company is also developing evaluation benchmarks that assess how robots handle complex social situations involving different personalities, preferences and relationships, as well as gesture-driven physical manipulation.
