Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

ML / AI Jobs in China

There are 2,373 open ML / AI jobs in China on freehire right now. 596 of them were posted recently. The skills employers ask for most often are python, cpp and pytorch.

Most requested skills

  • python 23%
  • cpp 19%
  • pytorch 19%
  • llm 18%
  • ai 17%
  • tensorflow 12%
  • nlp 11%
  • rag 8%

How the work is done

  • Onsite 13 · 1%
  • Remote 10 · 0%
  • Hybrid 8 · 0%

Visa sponsorship offered in 89% of the 18 postings that state a position on it.

Seniority

  • Senior 187
  • Intern 171
  • Junior 31
  • Lead 4
  • Staff 4
  • C-level 1

Who is hiring

  • 1000+ employees 134
  • 501-1000 employees 5
  • 11-50 employees 1
2.4K jobs
Alibaba Group
22 views 2 days ago

千问事业部-千问大模型Agentic算法专家-北京/杭州

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou

Alibaba's Qwen (千问) team is hiring an Agentic AI algorithm expert to define and build next-generation agent capabilities for the Qwen LLM — task planning, tool/API use, multi-step reasoning, long-horizon autonomous execution, agentic post-training (SFT/RLHF/DPO/GRPO), and evaluation systems — upgrading Qwen from answering questions to executing tasks for hundreds of millions of users.

Agentic AI AI API LLM NLP +1 skills
 Sign in to see your match
Alibaba Group
26 views 2 days ago

Token Foundry-模型自进化算法研究员-Qwen模型训练

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou

Researcher on Alibaba's Qwen team (Token Foundry) building self-evolving LLM algorithms: automated synthesis and curation of Agent RL/SFT training data, mining of pretraining/CPT data, and automated training/inference infra tuning with Verl, Megatron, SGLang, and vLLM. Based in Beijing or Hangzhou.

SGLang
Alibaba Group
30 views 2 days ago

Token Foundry-AI for Science/Engineering 算法研究员-Qwen模型训练

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou, Shanghai

Research role on Alibaba's Token Foundry team training Qwen models for AI in science and engineering: builds agent systems and benchmarks for biology/chemistry research, and applies self-evolving, long-horizon multi-agent methods to engineering problems (EDA, CAD, financial engineering, kernel tuning). Requires Python, strong engineering and research ability.
Alibaba Group
36 views 2 days ago

Token Foundry-AI for Engineering 大模型算法研究员-Qwen基础模型

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Shenzhen, Hangzhou, Shanghai

LLM algorithm expert on Alibaba's Qwen foundation-model team (Token Foundry), advancing Qwen's ability to do real engineering work — chip design (EDA), CAD/CAE, simulation, industrial automation — via self-evolving multi-agent loops, engineering tool-chain integration, domain data pipelines, and CPT→SFT→RL training at scale.

Agentic AI AI CAD CLI Data Pipelines +2 skills
 Sign in to see your match
Alibaba Group
29 views 2 days ago

Token Foundry-Computer Use Agent算法研究员-Qwen基础模型训练

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou, Shanghai

Researcher on Alibaba's Qwen foundation model team (Token Foundry) advancing GUI and hybrid Computer Use Agents across computer, mobile and browser environments — large-scale pre-training and RL on synthetic/real digital-environment trajectories, plus user preference and feedback modeling for training and evaluation.
Alibaba Group
27 views 2 days ago

Token Foundry-Cowork Agent算法研究员-Qwen基础模型训练

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou, Shanghai

Algorithm researcher on Alibaba's Qwen foundation model team, training LLMs to solve real-world agentic tasks: building reliable agent harnesses/environments, mining high-value industry data (finance, legal, healthcare, etc.), designing verifiers and reward systems, and running large-scale real-world agentic RL across CPT/SFT/RL pipelines.

Agentic AI
Alibaba Group
34 views 2 days ago

Token Foundry-视觉语言大模型算法研究员-Qwen

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou, Shanghai

Vision-language multimodal algorithm researcher on Alibaba's Qwen (Token Foundry) team: runs research and experiments on multimodal pre-training (data curation, scaling laws, distillation, long context), post-training (SFT/RLHF), RL for multimodal reasoning and agents, and unified understanding-generation models. Core stack: Python, PyTorch, Transformers.

RLHF
Alibaba Group
14 views 2 days ago

ATH-高级机器学习基础设施工程师/专家-AI制药

Open 17d APACSenior H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

1、分布式训练平台：多卡分布式训练/续训微调场景工程支撑 2、推理服务：模型推理服务化/批量生成任务队列与调度 3、算力调度：训练与推理任务资源编排/成本控制与排队 4、依赖与许可证合规：梳理并会签关键依赖商用授权 1、多卡分布式训练与推理服务工程经验能解决显存/通信/断点恢复 2、会用主流调度工具(Slurm/Ray/Kubernetes之一) 【核心技能】 1、续训与微调的断点与数据版本管理 2、算力成本与排队公平性平衡…

Slurm
Alibaba Group
45 views 2 days ago

千问事业部-健康智能体算法专家-杭州

New APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

At Alibaba's Qwen (Tongyi Qianwen) unit, research and build health-focused LLM agents that plan, call tools, retrieve evidence, and manage long-horizon tasks like health consultation, report interpretation, and medication management. Core work spans agentic RL (GRPO/PPO/RLVR), recursive self-improvement, long-term memory, and safety/evaluation, using PyTorch and large-scale distributed training.

Agentic AI LLM PyTorch RAG RLHF
Sign in to see your match
Alibaba Group
21 views 2 days ago

千问事业部-运动健康 Agent 算法工程师-杭州

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

Algorithm engineer in Alibaba's Qwen business unit building a sports & health AI Agent for wearables from 0 to 1: turning PPG/ECG/IMU sensor data into personalized health insights, running LLM post-training/alignment (SFT/RLHF/DPO/RL), building agent evaluation systems, and shipping an AI health coach across device and cloud. Core stack: LLMs, PyTorch, Transformers/DeepSpeed/Megatron.

Agentic AI AI DeepSpeed LLM PyTorch +3 skills
 Sign in to see your match
Alibaba Group
22 views 2 days ago

千问事业部-千问医疗对话大模型算法专家-杭州/北京

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing, Hangzhou

Algorithm expert on Alibaba's Qwen team building the Qwen medical dialogue LLM: develops multi-turn, multimodal (text/voice/video) medical conversation capabilities for health consultation, report reading and medication advice, with post-training/alignment (SFT, RLHF, agentic RL), RAG/agent search, and eval pipelines. Based in Hangzhou or Beijing.

Agentic AI LLM RAG RLHF
Sign in to see your match
Meituan
38 views 2 days ago

预测算法工程师

Open 17d APAC
China
Beijing

Prediction algorithm engineer on Meituan's bike-share battery-swap and asset management team: designs and ships forecasting models (battery consumption, swap value) and vehicle fault detection using ML, deep learning, time-series forecasting, computer vision, and multimodal LLMs, working closely with product and engineering.
Maersk
2 days ago

Associate AI/ML Engineer

New APACFull-timeJunior
China
Sichuan, Chengdu, 610041

Data AI/ML (Artificial Intelligence and Machine Learning) Engineering involves the use of algorithms and statistical models to enable systems to analyze data, learn patterns, and make data-driven predictions or…

AI Analytics Automation Big Data CI/CD +17 skills
 Sign in to see your match
Alibaba Group
23 views 3 days ago

高德-ABot-M具身操控大模型算法工程师/专家-视觉技术中心

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing

Algorithm engineer/researcher on Alibaba's AMap CV Lab building ABot-M, a vision-language-action (VLA) foundation model for embodied robot manipulation. Day to day: designing and training VLA/world models in PyTorch, building a robot data engine, and driving sim-to-real deployment on robot arms and humanoids.

AI GitHub LLM PyTorch
Sign in to see your match
Alibaba Group
30 views 3 days ago

阿里国际站/Alibaba.com-Agent后训练算法工程师/专家-Accio

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

Researches and productionizes post-training algorithms (SFT, reinforcement learning, distillation) for the LLMs behind Accio, Alibaba.com's native AI agent for global B2B trade. Day to day: build agent trajectory data pipelines, design rewards/verifiers, and run large-scale training experiments in Python and PyTorch.

Python PyTorch
Sign in to see your match
Alibaba Group
22 views 3 days ago

高德-资深具身运控算法工程师/专家-具身业务部

New APACSenior H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Beijing

Senior embodied motion-control algorithm engineer/expert on Alibaba's Amap (Gaode) embodied-business team in Beijing: builds next-generation general-purpose robot locomotion controllers using deep reinforcement learning — sim2real policy deployment, multi-modal motion switching, and IsaacGym/IsaacSim training — with C++/Python, ROS/ROS2, and real-robot control.
Alibaba Group
27 views 3 days ago

达摩院-医疗AI Agent 算法工程师-杭州

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

Algorithm engineer at Alibaba's DAMO Academy in Hangzhou building clinical/medical AI agents: developing planning, tool-use (MCP), multi-agent collaboration, and medical-grade RAG systems, plus agentic RL training (PPO/GRPO, PRM) and SFT/DPO/RLHF alignment to power disease screening, assisted diagnosis, and clinical decision support.

Agentic AI AI MCP Python PyTorch +2 skills
 Sign in to see your match
Alibaba Group
14 views 3 days ago

淘宝平台事业部-Agent应用算法工程师-智能服务Agent Lab

Open 17d APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

1. 负责淘宝商家智能客服系统设计与优化，包含harness设计以及优化，提升询单转化、降低转人工率； 2. 根据业务实际需求大模型进行post training（SFT/RLHF等）优化，提升大模型推理精度； 3. 参与自进化、memory等前沿技术方向的探索和落地； 4. 跟进Agent前沿技术趋势，推动技术方案从实验到落地。 1. 计算机、数学或统计学相关专业硕士及以上学历； 2.…

AI C++ RAG
Sign in to see your match
Alibaba Group
53 views 3 days ago

菜鸟驿站-多模态CV算法工程师-杭州-驿站算法部

New APAC H-1B sponsor history — This employer has had H-1B petitions approved by USCIS in the last five fiscal years, per the USCIS H-1B Employer Data Hub. The approval belongs to the employer and is not a commitment to sponsor this role.
China
Hangzhou

1. 负责菜鸟驿站核心视觉算法的研发与优化，覆盖图像矫正、目标检测、OCR、版面结构化（Layout）、关键信息抽取（KIE）等方向 2. 参与图像表示学习与垂类多模态大模型训练相关工作 3. 负责算法在 IoT 端侧设备上的部署与性能优化 4. 跟进 Automatic Prompt Optimization（APO）、多模态大模型幻觉检测等前沿技术在业务中的落地 5.…

GPU IoT OpenCV Python PyTorch +2 skills
 Sign in to see your match
ShengShu Technology
9 views 3 days ago

多模态算法工程师（参考生链路方向）

New APAC
China
Beijing

1、构建人物、物体、场景等参考库，通过检索、聚类和匹配，生产“参考素材—目标视频”训练数据。 2、结合图像／视频生成、编辑及可控合成技术，建设主体一致性、动作参考、运镜参考等数据生产链路。 3、推进链路自动化、质量筛选与规模化交付，并结合模型训练反馈持续优化数据效果。 1、熟练掌握 Python、PyTorch，具备计算机视觉或多模态算法研发经验。 2、在多模态检索、检测分割、目标跟踪、图像／视频生成等至少一个方向有实际项目经验。…

Python PyTorch
Sign in to see your match