Tech Lead / Principal Researcher - 异构计算与推理系统研究主管 上海 社招 全职 互联网 / 电子 / 网游 职位 ID：A231267 职位描述 【岗位职责】前沿探索与学术产出（80%）：硬件感知的大模型推理研究（Hardware-aware Inference）：针对无问芯穹独有的“M种模型xN种异构芯片”生态，研究下一代面向海量混合异构集群的分布式推理架构、极致调度算法与软硬协同设计（Co-design）。编译优化与算子极限探索：研究前沿的图编译技术、动态Shape优化、基于异构NPU/GPU的高效算子生成与极限内存优化（Memory-bound优化等）。顶会论文/专利发表：主导团队在 OSDI、SOSP、ASPLOS、MLSys、ISCA、HPCA 等计算机系统与体系结构顶级会议，以及 NeurIPS、ICLR 等AI顶会发表高质量学术论文，并布局底层技术核心专利。技术转化与赋能（20%）：开发并验证突破性的推理加速系统原型（如探索端到端延迟的极限压缩），为公司下一代Infra产品规划技术路线图。将验证成功的系统架构、核心调度算法或极致算子逻辑，交接给底层的无问芯穹Infra工程团队，指导工业级C++/CUDA/Triton代码落地。顶尖研究团队建设：建立并带领一支约5人规模的系统架构与异构计算精锐研究团队，指导团队成员攻坚“硬核”系统级科研难题。 职位要求 岗位要求】学术背景：计算机科学、体系结构、电子工程等相关专业博士学历；在系统软件、编译原理、并行计算或体系结构方向有深厚积累。硬性指标：以第一作者在 OSDI、ASPLOS、MLSys、ISCA、MICRO 等系统/体系结构顶会，或NeurIPS/ICLR等AI顶会发表过高影响力论文。技术功底：深入理解大模型底层计算逻辑与并行策略（Tensor Parallel, Pipeline Parallel等）。熟悉主流推理框架（vLLM, TensorRT-LLM, Megatron等）及底层编译工具链（Triton, TVM, MLIR等）。具备极强的C++/CUDA/底层汇编级优化理解力，能手撕极致性能的算子原型。团队领导力：有带领系统方向科研团队的经验，热衷于解决极具挑战性的工程与科学交叉难题。 职位信息 部门： 智能终端产研
About Us SharkNinja is a global product design and technology company, with a diversified portfolio of 5-star rated lifestyle solutions that positively impact people’s lives in homes around the world. Powered by two trusted, global brands,
NVIDIA has been redefining computer graphics, PC gaming, and accelerated computing for more than 30 years. It’s a unique legacy of innovation that’s fueled by great technology—and amazing people. Today, we’re tapping into the unlimited potential
NVIDIA has been transforming computer graphics, PC gaming, and accelerated computing for more than 25 years. It’s a unique legacy of innovation that’s fueled by great technology—and amazing people. Today, we’re tapping into the unlimited potential
后训练桌面端工程师 北京、上海 社招 全职 互联网 / 电子 / 网游 职位描述 岗位背景我们在为大模型做「移动 & 桌面 & Web 应用开发」方向的后训练(SFT / RL / Agent 评测)。桌面端是当前模型能力最薄、但用户最在意的一块 —— Windows / macOS / Linux 三个平台、Electron / Tauri / Qt / 原生框架五花八门、还要跟图形栈(wgpu / Skia
About VAT Group At VAT, we change the world with vacuum solutions. As the world’s leading supplier of high-performance vacuum valves, we have been driving innovation for more than 60 years. With over 3,200 employees worldwide,
Intelligent machines powered by Artificial Intelligence computers that can learn, reason and interact with people are no longer science fiction. GPU Deep Learning has provided the foundation for machines to learn, perceive, reason and solve problems.
具身算法引擎开发实习生 上海 实习 职位描述 1. 训练框架开发：参与VLA等大模型训练框架研发与优化，包括数据并行、模型并行、流水并行等分布式训练能力的建设，提升千卡规模的训练稳定性与资源利用率。2. 推理系统优化：参与VLA等大模型的推理引擎开发与性能优化，包括算子优化、KV Cache管理、Prefill/Decode优化、Batch调度、低延迟Serving等核心能力建设，提升推理吞吐与响应速度。3. 分布式系统优化：参与多机多卡训练与推理系统建设，优化NCCL通信、AllReduce效率及IB/RDMA网络性能，提升大规模集群运行效率。4. 稳定性保障：解决大规模训练中的故障恢复、Checkpoint 管理、梯度异常等问题，确保长时间训练任务的稳定运行。5. 平台工程建设：参与训练/推理平台工程化建设，包括模型部署、监控、自动化测试、性能Benchmark及稳定性保障。 职位要求 1. 本科及以上学历，计算机、人工智能、自动化、机器人等相关专业优先2. 熟悉Python或C++开发，具备良好的代码能力与工程习惯3. 精通 PyTorch 分布式训练机制（DDP/FSDP），熟悉 DeepSpeed、Megatron-LM 、vLLM、SGLang等大规模训练/推理框架的原理与使用4. 深入理解 3D 并行（Data Parallel / Tensor Parallel / Pipeline Parallel）的实现原理与适用场景，能够根据模型特点设计最优并行策略5. 熟练使用 PyTorch Profiler、NVIDIA Nsight 等工具进行性能分析，能够定位并解决计算、通信、I/O 瓶颈加分项：1. 有 LLM/VLM/VLA
Company Description 该岗位现面向所有经验阶段的候选人开放，包括社会招聘、应届毕业生，同时开放实习生岗位。工作地点为北京。欢迎申请，期待你的加入！ Notice: This position is open to candidates at all experience levels, including experienced candidates, graduates, as well as internship opportunities. The role is based in Beijing. We welcome your application and look forward
Job Description Work Arrangement This role is categorized as on-site. The successful candidate is expected to report to the Advanced Technical Center full time. The Role - The Interior Design Manager is a senior leadership role
Job Description SummaryAs a Production Sourcing Leader, you will play a critical role in supporting the Magnetic Resonance Imaging (MRI) business by leading high-impact sourcing initiatives focused on cost optimization, productivity, and supply chain resilience. This
The Buyer – Trimmings plays a pivotal role in Suitsupply’s supply chain, leading the end-to-end sourcing and purchasing of trimmings across a network of approximately 45 suppliers. Based in Shanghai and reporting to the Buying Lead
Job Purpose The Test Engineer is responsible for communicating with a client that drives forward User Acceptance Testing to participate in their functional and regression testing using their internal front and back end systems, apps and
At Gilead, we’re creating a healthier world for all people. For more than 35 years, we’ve tackled diseases such as HIV, viral hepatitis, COVID-19 and cancer – working relentlessly to develop therapies that help improve lives
Intelligent machines powered by AI computers that can learn, reason and interact with people are no longer science fiction. Today, a self-driving car can meander through a country road at night and find its way. An
About impact.com impact.com is the world’s leading commerce partnership marketing platform, transforming the way businesses grow by enabling them to discover, manage, and scale partnerships across the entire customer journey. From affiliates and influencers to content
Job TitleSr. Supplier Development Engineer Job Description We are seeking an experienced Lead Auditor to conduct external supplier audits for medical device components, products, and services. This role is critical in ensuring that our supply base
NVIDIA is leading company of AI computing. At NVIDIA, our employees are passionate about AI, HPC , VISUAL, GAMING. Our SA team is more focusing to bring NVIDIA new technology into difference industries. We help to
Position Summary Responsible for electrical system design of supercapacitor energy storage systems, covering single-cabinet to multi-cabinet parallel configurations. Candidates must demonstrate independent design capability for main circuit topology and protection circuits, deep understanding of energy exchange interfaces
关于我们 / About Us AlphaGrep 是一家全球领先的量化交易公司，专注于股票、商品、外汇及固定收益等资产的算法交易。我们在国际市场拥有显著份额，依托自主开发的超低延迟系统与严格的风控体系，持续构建高效能策略。 AlphaGrep is a leading global quantitative trading firm specializing in algorithmic strategies across equities, commodities, FX, and fixed income. We hold significant market share internationally, powered by proprietary low-latency infrastructure