算法应用工程师（具身智能VLA操作大模型方向）
Tencent
算法应用工程师（具身智能VLA操作大模型方向）
Job Description
算法应用工程师（具身智能VLA操作大模型方向）
📋 Job Overview
The Algorithm Application Engineer role in Embodied Intelligence VLA Operation Large Model direction at Tencent focuses on optimizing and deploying VLA models for robot operation tasks. Responsibilities include performance enhancement through evaluation and iteration cycles, hardware integration, and cross-team collaboration to ensure reliable system deployment. The position involves surveying open-source models, data management, and advancing practical applications in robotics.
📍 Location: Shenzhen, China
🏢 Business Unit: TEG
📄 Full Description
2.设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地；
3.在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠；
4.调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀；
5.数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代；
6.跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全。
🎯 Key Responsibilities
- 基于现有具身智能VLA操作大模型，深入理解模型架构、训练范式与推理链路，针对具体机器人操作任务进行性能优化（如成功率、鲁棒性、时延），形成“评测—分析—优化—回归”闭环，稳定提升落地效果
- 设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地
- 在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠
- 调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀
- 数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代
- 跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全
🛠️ Required Skills
- Deep understanding of VLA model architecture, training paradigms, and inference pipelines
- Experience in robot operation task optimization (success rate, robustness, latency)
- Skills in model evaluation, dataset construction, and bottleneck analysis
- Hardware integration with robots and mechanical arms (sensors, actuators, hand-eye calibration, time synchronization)
- Fault diagnosis and recovery strategies for hardware anomalies
- Surveying and reproducing open-source VLA models (e.g., RT series, OpenVLA, Open-X-Embodiment)
- Data collection, cleaning, augmentation, and resampling
- Cross-team collaboration with software, hardware, testing, and production teams
Locations
- Shenzhen, China
Salary
300,000 - 800,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- Deep understanding of VLA model architecture, training paradigms, and inference pipelinesintermediate
- Experience in robot operation task optimization (success rate, robustness, latency)intermediate
- Skills in model evaluation, dataset construction, and bottleneck analysisintermediate
- Hardware integration with robots and mechanical arms (sensors, actuators, hand-eye calibration, time synchronization)intermediate
- Fault diagnosis and recovery strategies for hardware anomaliesintermediate
- Surveying and reproducing open-source VLA models (e.g., RT series, OpenVLA, Open-X-Embodiment)intermediate
- Data collection, cleaning, augmentation, and resamplingintermediate
- Cross-team collaboration with software, hardware, testing, and production teamsintermediate
Responsibilities
- 基于现有具身智能VLA操作大模型，深入理解模型架构、训练范式与推理链路，针对具体机器人操作任务进行性能优化（如成功率、鲁棒性、时延），形成“评测—分析—优化—回归”闭环，稳定提升落地效果
- 设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地
- 在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠
- 调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀
- 数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代
- 跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全
Tags & Categories
Answer 10 quick questions to check your fit for 算法应用工程师（具身智能VLA操作大模型方向） @ Tencent.
Related Books and Jobs
算法应用工程师（具身智能VLA操作大模型方向）
Tencent
算法应用工程师（具身智能VLA操作大模型方向）
Job Description
算法应用工程师（具身智能VLA操作大模型方向）
📋 Job Overview
The Algorithm Application Engineer role in Embodied Intelligence VLA Operation Large Model direction at Tencent focuses on optimizing and deploying VLA models for robot operation tasks. Responsibilities include performance enhancement through evaluation and iteration cycles, hardware integration, and cross-team collaboration to ensure reliable system deployment. The position involves surveying open-source models, data management, and advancing practical applications in robotics.
📍 Location: Shenzhen, China
🏢 Business Unit: TEG
📄 Full Description
2.设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地；
3.在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠；
4.调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀；
5.数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代；
6.跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全。
🎯 Key Responsibilities
- 基于现有具身智能VLA操作大模型，深入理解模型架构、训练范式与推理链路，针对具体机器人操作任务进行性能优化（如成功率、鲁棒性、时延），形成“评测—分析—优化—回归”闭环，稳定提升落地效果
- 设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地
- 在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠
- 调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀
- 数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代
- 跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全
🛠️ Required Skills
- Deep understanding of VLA model architecture, training paradigms, and inference pipelines
- Experience in robot operation task optimization (success rate, robustness, latency)
- Skills in model evaluation, dataset construction, and bottleneck analysis
- Hardware integration with robots and mechanical arms (sensors, actuators, hand-eye calibration, time synchronization)
- Fault diagnosis and recovery strategies for hardware anomalies
- Surveying and reproducing open-source VLA models (e.g., RT series, OpenVLA, Open-X-Embodiment)
- Data collection, cleaning, augmentation, and resampling
- Cross-team collaboration with software, hardware, testing, and production teams
Locations
- Shenzhen, China
Salary
300,000 - 800,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- Deep understanding of VLA model architecture, training paradigms, and inference pipelinesintermediate
- Experience in robot operation task optimization (success rate, robustness, latency)intermediate
- Skills in model evaluation, dataset construction, and bottleneck analysisintermediate
- Hardware integration with robots and mechanical arms (sensors, actuators, hand-eye calibration, time synchronization)intermediate
- Fault diagnosis and recovery strategies for hardware anomaliesintermediate
- Surveying and reproducing open-source VLA models (e.g., RT series, OpenVLA, Open-X-Embodiment)intermediate
- Data collection, cleaning, augmentation, and resamplingintermediate
- Cross-team collaboration with software, hardware, testing, and production teamsintermediate
Responsibilities
- 基于现有具身智能VLA操作大模型，深入理解模型架构、训练范式与推理链路，针对具体机器人操作任务进行性能优化（如成功率、鲁棒性、时延），形成“评测—分析—优化—回归”闭环，稳定提升落地效果
- 设计并实施模型评测与实验方案：构建任务集/数据集、定义核心指标、进行评测，按阶段复盘，定位瓶颈，输出可实施的优化方案并推进落地
- 在真实机器人与机械臂硬件上完成集成与验证：开展传感器与执行器接入、手眼标定与时间同步、故障定位与恢复策略，能分析并处理硬件侧异常现象，确保系统整体闭环可靠
- 调研、导入开源VLA操作大模型：快速理解并复现开源方案（如RT系列、OpenVLA、Open-X-Embodiment等），修复适配问题，完成横向竞品性能对比与技术报告沉淀
- 数据闭环与场景覆盖：参与数据采集、清洗、增强与重采样，完善关键任务场景与长尾问题的覆盖，支持线上/线下迭代
- 跨团队协作：与软件、硬件、测试及生产团队紧密配合，制定落地方案与风险控制策略，保障上线质量与安全
Tags & Categories
Answer 10 quick questions to check your fit for 算法应用工程师（具身智能VLA操作大模型方向） @ Tencent.
Related Books and Jobs