Mbodi, a New York-based startup, aims to revolutionize robot training using AI agents, making the process faster and simpler. As a finalist in TechCrunch Disrupt 2025, Mbodi showcases its innovative cloud-to-edge hybrid computing system designed to integrate seamlessly into existing robotic technologies. Co-founders Xavier Chi and Sebastian Peralta recognized the limitations in traditional robot training while working at Google. Their system employs a network of AI agents that work collaboratively, breaking down tasks into manageable subtasks, allowing robots to learn efficiently from real-world data.
While many companies are focused on building extensive AI models for training, Mbodi addresses the need for adaptability in constantly changing environments. Currently, they are collaborating with a Fortune 100 consumer goods company to streamline packaging processes, which face daily variations. Mbodi aims to begin broader software deployment by 2026, emphasizing practical solutions over theoretical research. For more insights, attendees can join the TechCrunch Disrupt event in San Francisco from October 27-29, 2025.
Source link
