1. Participate in the R&D and optimization of edge and cloud-side inference systems for high-performance embodied model execution.
2. Design high-performance custom operators and efficient quantization/compression schemes to maximize hardware theoretical performance.
3. Lead the architectural design of embodied AI inference engines; write high-quality C++/Python code and build automated performance testing and monitoring systems.
4. Deeply involved in inference acceleration and model miniaturization for large-scale models.

1. Participate in the development of training infrastructure for large-scale embodied AI models.
2. Iterate on training framework architecture to enhance stability, usability, maintainability, and observability.
3. Support performance optimization for large-scale training systems with thousands of GPUs.

1. Design and iterate on annotation and pre-annotation schemes for embodied AI pre-training data; develop corresponding pre-annotation models to improve data production efficiency.
2. Build data quality assessment and quality control models tailored to embodied data collection and usage needs, enabling a closed-loop data quality process.
3. Collaborate closely with algorithm, data, and business teams to continuously optimize data workflows, providing high-quality data to support model training and performance improvement.

1. Conduct pre-training and research on foundation models tailored for humanoid robots; explore model architectures better suited for embodied tasks.
2. Explore multimodal (Vision-Language-Action) representation alignment and fusion to provide robust feature support for high-level decision-making.
3. Drive the migration and application of vision foundation models to robotic downstream tasks such as navigation and manipulation.

- Lead the design of computer vision algorithms for humanoid robots, covering model architecture, training, deployment, and fine-tuning.
- Manage data organization, analysis, and mining for relevant business areas to enhance model accuracy and generalization.

- Develop core machine learning systems for embodied AI.
- Develop VLA training and reasoning systems for robotics scenarios.
- Support the continuous evolution of multi-modal perception, motion control, and task planning systems.
- Develop toolchains for embodied AI systems, covering the full lifecycle of data collection, simulation training, physical deployment, and continuous optimization.

- Build a machine learning platform tailored for embodied AI, supporting the training and evaluation of VLA models.
- Develop an intelligent data factory platform to manage the full lifecycle of data collection task dispatch, personnel/equipment scheduling, data annotation, and quality monitoring.
- Design a cross-cloud, edge, and device computing architecture to ensure high performance and reliability in complex scenarios.
- Optimize data storage and processing workflows, establishing a positive feedback loop between data quality and model performance.
- Collaborate with robotics and algorithm teams to drive the transition of embodied AI from theoretical research to practical applications.

- Develop and optimize the engineering implementation of VLA.
- Design and implement post-processing fusion algorithms for embodied VLA models, integrating information from cameras, motors, and other sensors to detect abnormal conditions and ensure stable system operation.
- Participate in the architectural design of embodied intelligence fusion algorithms and resolve key technical issues.
- Deploy and debug relevant algorithms on actual devices, and promote business application implementation.
5. Continuously optimize and iterate the implementation and performance of embodied intelligence algorithms.

1. Responsible for VR and isomorphic exoskeleton teleoperation system design and algorithm development, including master-slave mapping, VR video streaming, wireless/remote operation, and force feedback.
2. Collaborate with software, hardware, and AI teams to develop and deliver teleoperation data acquisition systems/products, systematically optimizing performance (latency, frame loss, jitter) to enhance data collection quality and efficiency for humanoid robots.
3. Oversee the full lifecycle (requirements, specs, design, prototyping, testing, production) of teleoperation systems; lead systematic analysis and resolution of issues encountered.
4. Track global frontiers in teleoperation and data acquisition to empower the company's data-driven product ecosystem.

1. Collaborate with related modules to develop and debug force-control algorithms (e.g., drag-to-teach, admittance/impedance control, collision detection, force feedback, and hybrid force/position control).
2. Produce design documentation for force-control software and algorithms; formulate comprehensive test plans and test cases.
3. Handle motion control tasks, including trajectory planning, forward/inverse kinematics and dynamics modeling, dynamics feedforward/feedback control, system identification, and parameter tuning.

1. Lead the hardware implementation for wearable data-acquisition gear and robot bodies, including solution design, circuit design, and electrical component selection.
2. Responsible for embedded circuit design, key component selection, schematic and PCB design, test verification, and New Product Introduction (NPI).
3. Collaborate with software, structural, and wiring harness engineers to complete full-system design and debugging.
4. Optimize system stability, reliability, and maintainability; support mass production and rapid iteration while resolving issues during the development-to-production lifecycle.
5. Monitor global trends in data-acquisition equipment and innovate product designs to ensure long-term competitiveness.

1. Responsible for Linux BSP development and maintenance on AI hardware platforms, including Bootloader porting/optimization, kernel stripping/configuration, and Device Tree coding/debugging.
2. Collaborate with vendors on driver development for buses (I2C, SPI, UART, CAN, USB, Ethernet, MIPI) and peripherals (sensors, motors, GMSL cameras).
3. Collaborate with hardware engineers on board-level hardware-software debugging and systematic fault isolation and resolution.
4. Perform system optimization to resolve driver-level issues regarding real-time performance, power consumption, and compatibility to maximize efficiency.
5. Compose driver development documentation, test cases, and maintenance manuals.
6. Establish the functional and performance evaluation framework for chips/SoCs.

1. Responsible for the detailed design of robotics products; deeply involved in core technology development and verification, including component selection, overall layout, and module partitioning.
2. Contribute to mechanical team standardization, skill enhancement, and technical pre-research; promote the documentation and reuse of team expertise and technologies.
3. Optimize product quality, cost, and schedule; drive improvements in Design for Testing (DFT) and Design for Manufacturing (DFM).
4. Complete detailed 3D/2D CAD designs to ensure performance and manufacturability; lead structural research, design, and innovation for new technology projects.

- Responsible for the overall design of the robot and base, lead the development and verification of core technologies, and complete tasks such as dimensioning and verification of key components, overall layout, and module division.
- Responsible for standardizing the work of the mechanical team, enhancing skills, and conducting technical pre-research. Lead the resolution of key technical issues and promote summarization and repeated application of team experience and technology.
- Responsible for improving the quality, cost, and schedule of product sections, and promoting the improvement of product testability and manufacturability.
- Responsible for completing detailed 3D and 2D drawing designs, lead the organization of scheme and drawing reviews, and ensure the product's performance and good production process.
- Responsible for the structural research, design, and innovation of new technology projects, and ensure the innovation and leading position of technology.

- Design, development, and delivery of humanoid robot hardware systems.
- Design of hardware solutions for robot perception systems, selection and introduction of key components, including but not limited to cameras, LiDARs, IMUs, etc.
- Delivery of hardware platform and project-related deliverables: plans, designs, hardware architecture, network topology, wiring layout, electrical energy management, etc.
- Review of hardware and electrical technical solutions, pre-estimation of technical difficulties and risk identification in system solutions, inter-departmental cooperation and technical support.
- Competitive research on humanoid robot hardware and electrical aspects, and mid-to-long-term technical planning.

- Research and implementation of algorithms related to path generation and trajectory planning of humanoid robot arms, modeling identification and compensation, dynamics and control.
- Proactively identify positioning issues and propose technical solutions, and suggest improvements for collaborative design between the robot's mechanical structure and algorithms based on its motion characteristics.
- Participate in the construction and simulation of robot kinematic and dynamic models, and establish relevant standards for robot motion performance.
- Participate in the development of one or more cutting-edge technologies, including but not limited to planning and control in the fields of robot posture, multi-arm collaboration, and structural reconfiguration; conduct in-depth review and summarization of technical literature in these fields.

- Responsible for the development of autonomous navigation functions for wheeled humanoid robots, including the implementation of algorithms for path planning, obstacle avoidance, and escape from entrapment.
- Design an efficient planning and navigation framework, optimize algorithm performance, ensure engineering implementation, and continuously iterate.
- Collaborate with SLAM, chassis hardware, and motion control teams to analyze and close the loop on issues that arise during robot operation.
- Track cutting-edge planning algorithms, including but not limited to end-to-end AI navigation, such as GOAT (go to anything).

- Participate in the software development of robot controllers (including algorithm integration, communication middleware, bus drivers, etc.), follow the software development process, and be able to independently design and implement software.
- Collaborate with colleagues in algorithm, hardware, and AI to promote the system integration and ecosystem development of humanoid robot products.
- Participate in the analysis and resolution of issues related to the performance and reliability of the robot control system.
- Work with team members to build up software architecture, enhancing software development efficiency and quality.