AgiBot
Posted 1y ago

具身算法专家/工程师

AgiBot
Beijing or Shanghai
OnsiteFull Time
Responsibilities
  • designing algorithms
  • developing algorithms
  • optimizing models
Requirements
  • Design and develop imitation and reinforcement learning algorithms for humanoid robots
  • Build multimodal perceptual and decision models
  • Strong Python and deep learning framework experience
  • Master\u0002/PhD in related fields required
Technical tools mentioned
PythonPyTorchJAXTensorFlowPaLM-ER3MRFMOpenVLA

Job description

1.设计和开发创新性的模仿学习和强化学习算法,应用于人形机器人场景任务,具备泛化能力
2.研究并构建具备视觉、触觉、语言等多模态感知和决策能力的大模型,实现机器人在开放世界中的物理交互。
3.优化具身算法模型结构,以支持具身算法的实现和高效运行。
4.与跨职能团队合作,推动具身算法在实际场景中的应用

1.计算机视觉、多模态大模型等相关领域的硕士或博士
2.精通Python编程语言,熟悉PyTorch、JAX、TensorFlow等深度学习框架
3.熟悉机器人模仿学习、强化学习理论,有相关领域有影响力的Paper者优先
4.有ACT、MT-ACT、RT系列、Diffusion Policy,GO/GE World系列,PI系列,GR00T系列,HIL-SERL等算法相关研究背景经验者优先
5.熟悉LLM/VLM/VLA,了解R3M、RFM、PaLM-E、OpenVLA,RL等多模态具身大模型经验者优先
6.有实际的机器人、自动驾驶项目经验者优先。

About AgiBot

Developing embodied AI and humanoid robotics technology.

Similar jobs

Embodied Algorithm Engineer roles near Beijing, Beijing
1mo
Save
Mark Applied
Hide
具身算法工程师-Genie业务部
Shanghai or Beijing
OnsiteFull Time
Agibot
Agibot: Develops general-purpose humanoid robots and embodied AI systems.
Master's or PhD in computer vision or multimodal large models; strong Python and deep learning background; experience in imitation learning, reinforcement learning, multimodal embodied models and robot projects preferred.
Python, PyTorch, JAX, TensorFlow
4mo
Save
Mark Applied
Hide
具身算法工程师(多模态与System0方向)
Beijing, Beijing, China
OnsiteFull Time
Dexmal
Dexmal: Develops end-to-end embodied AI software and robotics hardware.
Design multimodal high-frequency control models using large-scale training and RL; strong C++ and Python skills, strong experimental ability; robot manipulation experience preferred.
C++, Python
1y
Save
Mark Applied
Hide
小米汽车-具身算法工程师/专家
Beijing, Beijing, China
OnsiteFull Time
Xiaomi
XiaomiHong Kong Stock Exchange: 1810: Designs and manufactures smartphones, consumer electronics, and smart hardware.
Master's degree or above in robotics/CS/AI, strong in imitation/ reinforcement learning, PyTorch or TensorFlow, robot simulators and ROS, C++/Python programming, model training and deployment experience.
PyTorch, TensorFlow, Octo, OpenVLA, GR-2, Mujoco, Isaac Lab, ROS, C++, Python
1mo
Save
Mark Applied
Hide
具身智能算法工程师
Beijing, Beijing, China
OnsiteFull Time
Beijing Humanoid Robot Innovation Center
Beijing Humanoid Robot Innovation Center: Develops humanoid robots and embodied intelligence AI platforms.
Master's or higher in CS/AI/robotics, experience in robot manipulation algorithms and VLM/VLA work, strong Python and PyTorch skills, docker and git familiarity, real-robot debugging or simulation experience.
Python, PyTorch, docker, git, Issac Lab, Mujoco
1mo
Save
Mark Applied
Hide
具身智能算法工程师(框架方向)
Beijing or Shanghai
OnsiteFull Time
Infinigence
Infinigence: AI infrastructure provider optimizing heterogeneous computing for large models.
1+ YOEMaster's degree or higher in computer science, robotics, automation, or artificial intelligence; 1+ year of reinforcement or robot learning experience; RL algorithms, model fine-tuning, distributed training, and framework development experience.
RLinf, PPO, SAC, LoRA, QLoRA, VLA, WAM, VLM, LLM, Checkpoint
1mo
Save
Mark Applied
Hide
具身智能算法工程师-世界模型与长程任务
Beijing, Beijing, China
OnsiteFull Time
SenseTime
SenseTimeHKEX: 0020: Develops artificial intelligence software and computer vision technology.
Master's degree (2027) or above in CS/AI/Robotics/Automation/Electrical; strong ML/deep learning and robotics foundation; PyTorch experience; English reading and collaboration skills; able to design experiments and deploy on robot/simulation.
PyTorch
1mo
Save
Mark Applied
Hide
【27届校招】具身智能后训练算法工程师
Beijing, Beijing, China
OnsiteFull Time
Dexmal
Dexmal: A research and development building embodied intelligence and robotic capabilities for real-world tasks.
Bachelor's degree in CS/AI/robotics/electronic engineering, strong experience with LLM/VLM/VLA, world models, PyTorch, distributed training, Python/C++, sim-to-real and real-robot deployment.
LLM, VLM, VLA, World Model, World Action Model, PyTorch, DDP, FSDP, DeepSpeed, Ray, Python, C++, ROS, ROS2, Isaac Sim, Mujoco, PyBullet, ManiSkill, RoboSuite, OpenVLA, RT-1, RT-2, ACT, Diffusion Policy, Octo, SayCan, PaLM-E
2mo
Save
Mark Applied
Hide
具身智能算法工程师(World Action Models策略方向)
Beijing, Beijing, China
OnsiteFull Time
Shengshu Technology
Shengshu Technology: Develops multimodal generative AI models and world model frameworks.
1+ YOEBachelor's in CS/automation/robotics/AI,1+ years in embodied intelligence or robot strategy algorithms; expertise in world models, action policies, Diffusion Policy, Transformer, RL, imitation learning; PyTorch and ROS/ROS2 experience; simulation and real-robot deployment experience.
PyTorch, ROS, ROS2, Isaac, Sapien, Transformer, Diffusion Policy