Long Chen

Chen is committed to incorporating the reasoning and language capabilities of Large Language Models (LLMs) into autonomous driving systems, aiming to move beyond the limitations of traditional end-to-end AI and build a "transparent brain" that can not only drive, but also articulate actions and intentions in natural language. The core of this work is to enable vehicles to clearly explain their reasoning, making their decision-making process understandable and auditable to humans.

To achieve this, he and his team at Wayve developed the "Lingo" series, one of the first Vision-Language-Action (VLA) model frameworks. The precursor "Driving-with-LLMs" work in 2023 enables an autonomous driving system to understand complex traffic scenarios by fusing multimodal information from vehicle perception, particularly object-level vector modality. He and his team further expanded this framework, developing the "LingoQA" model to answer complex questions about real-world driving scenes and "Lingo-2" model, the first closed-loop VLA driving model tested on public roads.

Long Chen's research offers a new paradigm for next-generation autonomous systems: a shift from purely data-driven executors to trustworthy intelligent agents capable of effective communication and common-sense reasoning like humans. Such intelligent systems have the potential to fundamentally enhance public trust in autonomous driving technology, accelerating its safe deployment and application in the real world. After joining Xiaomi in 2025, he has continued to lead the development of Embodied AI, bringing advanced cognitive capabilities into the physical world by building foundational embodied agents in autonomous vehicles, robots, and smart devices that can perceive, reason, and act in complex environments, taking a monumental step toward Physical Artificial General Intelligence (Physical AGI).