AI Infra Engineer(Agentic Runtime)
深圳、北京、上海
全职
智能机器人板块
职位描述
参与从 0 到 1 建设具身学习基础设施平台,为机器人大模型的强化学习后训练、评测与自迭代提供统一的环境服务与策略推理能力。以仿真环境和策略推理为切入点,通过资源调度、系统性能优化和运行保障,提高有效样本产出,降低实验等待与资源成本,并逐步扩展至端侧和真机执行场景。 【职位描述】 1. 环境与 Rollout 服务平台建设:建设共享 Sandbox 资源池与执行服务,实现环境申请、初始化、隔离、重置、回收与运行观测;对接强化学习训练框架,支撑真实具身模型训练与评测任务; 2. 策略推理集成与优化:集成和优化 vLLM、SGLang 等推理引擎,建设策略服务部署、路由与资源管理能力;结合业务负载优化批处理、缓存、多模态处理与模型执行效率,保障权重更新与模型版本关联的正确性; 3. 资源画像与动态调度:根据任务类型、环境资产和实际资源用量建立资源画像,设计预热、资产亲和、配额管理与动态扩缩容策略,通过运行反馈持续优化资源申请规格、并发配置与调度效率; 4. 端到端性能分析:分析环境准备、仿真执行、观测传输、策略推理及权重同步等阶段的瓶颈,建立可复现的性能基线,以有效 Rollout 吞吐、端到端时延和资源成本验证优化收益; 5. 稳定性与执行正确性保障:完善故障检测、超时处理、背压、恢复与资源清理机制,支持长时间无人值守运行;建立任务、环境、轨迹、Reward 与 Policy 版本的关联追溯,保障执行隔离与样本正确性; 6. 标准化接入与架构演进:设计可扩展的任务与 Runtime 接口,支持 Isaac Sim、Unity 等仿真环境,以及强化训练、评测、Replay 等业务复用平台能力,为后续边缘节点、台架和机器人接入奠定基础。
职位要求
投递

XPeng is a leading Chinese Smart EV company that designs, develops, manufactures, and markets Smart EVs that appeal to the large and growing base of technology-savvy middle-class consumers. Its mission is to drive Smart EV transformation with technology and data, shaping the mobility experience of the future. In order to optimize its customers’ mobility experience, XPeng develops in-house its full-stack advanced driver-assistance system technology and in-car intelligent operating system, as well as core vehicle systems including powertrain and the electrical/electronic architecture. XPeng is headquartered in Guangzhou, China. In 2021, the Company established its European headquarters in Amsterdam, along with other dedicated offices in Copenhagen, Munich, Oslo, and Stockholm.The Company’s Smart EVs are mainly manufactured at its plant in Zhaoqing and Guangzhou,Guangdong province.
For more information, please visit https://heyxpeng.com.