Rohit John Varghese is the director of systems engineering and product at Contoro Robotics, a company developing AI-driven logistics robots for shipping container unloading. He has been part of the founding team of two successful startups that together have raised over $25 million in funding. His work spans exoskeletons, human–robot interaction, teleoperation, and AI-enabled robotic systems, and he is always curious to explore any domain to its first principles.
Rohit has served as faculty at The University of Texas at Austin, where he taught graduate courses in robotics, and as an industry mentor for National Science Foundation (NSF) programs. He also leads the development of the Robot MCP server, an open-source project that connects AI models with robotic systems using Model Context Protocol.
Model Context Protocol (MCP) has emerged as a standard for connecting AI systems with external tools, opening new pathways for collaboration between reasoning models and the physical world. The open-source Robot MCP project extends this standard to robotics—enabling AI models such as Claude, ChatGPT, and Gemini to perceive, plan, and act through existing robot ecosystems built on platforms like ROS.
This talk explores how the Robot MCP ecosystem is expanding through academic and industry partnerships to define open best practices, safety frameworks, and shared tool libraries. We’ll showcase how multi-agent orchestration—where perception, reasoning, and control agents coordinate through MCP—unlocks new forms of robotic autonomy and problem-solving.
By connecting the broader AI and robotics communities through a unified protocol, Robot MCP is shaping a future where intelligent agents and physical systems collaborate seamlessly, safely, and openly.