XPeng’s AI infrastructure head, Lu Siyuan, is leaving the company and heading to OpenAI to work on embodied intelligence robots. He is currently transitioning his responsibilities.
Lu oversaw the entire AI infrastructure pipeline — from model training to production deployment in vehicles — managing a team of about 200 people. His work covered training frameworks, GPU clusters, compilers for XPeng’s self-developed chips, model quantization optimization, and on-vehicle deployment.
This role was especially critical for XPeng, which relies on its own chip designs and therefore builds its own compilers and inference runtime systems. Because a replacement has not yet been found, Lu’s team is expected to be split into three or more independent groups.
OpenAI’s website is still hiring for robot software, simulation, and firmware engineers. The company says its robotics team is developing general-purpose robots that can work in real-world environments, bridging models, hardware, and software.