特斯拉Staff/Sr. Data Algorithm Engineer-CV

任职要求

必备条件
具备计算机科学、数学、统计学或相关学科的高等学历。
具备较强的C++和Python开发背景，能够处理高性能3D视觉算法实现。
具备很强的3D视觉算法背景，包括点云滤波、分割、检测、配准和6D位姿估计等，以及视觉尺寸测量算法（如3D重建和计量技术）
具备TensorFlow或PyTorch等框架的模型开发经验，并理解常用神经网络（如卷积网络、点云网络）在3D应用中的原理和优化。
具备数据科学相关能力，如Pandas、Numpy、Matplotlib以及MongoDB Aggregation等工具的使用经验。
了解机器人相关控制技术（如ROS框架、路径规划），并具备工厂3D机器视觉经验，包括视觉尺寸测量在制造质量控制中的应用。

优先考虑
有实际将3D视觉技术应用于工业、制造或机器人领域的项目经验，例如自动引导系统或缺陷检测，或视觉尺寸测量系统。
在相关领域的学术期刊或会议（如CVPR、ICCV、IROS）上发表过研究论文。
具有机器人/PLC控制、工业相机/激光传感器/光源解决方案的经验。
有在敏捷环境中工作的经验。
具有算法开发竞赛经验，如ACM、Kaggle或Robotics挑战赛。
熟悉多模态数据融合（如RGB-D与点云结合）或实时3D重建技术，特别是在视觉尺寸测量中的应用（如融合传感器数据提升测量精度）。

加入特斯拉，您将有机会在一个充满活力和创新的环境中工作，与世界一流的工程师和科学家一起，通过机器视觉技术推动工业自动化和智能制造的发展。如果您对机器学习、人工智能和机器视觉充满热情，并渴望在这一前沿领域中实现自己的价值，欢迎加入我们的团队。

The Role
The Data Algorithms team plays a key role in Tesla's industrial intelligence research and development. We build our own data algorithms platform to empower areas such as production manufacturing, supply chain, sales, services, and charging networks, transforming information into high-value data assets to create superior products and deliver a perfect user experience.

As a Data Algorithms Engineer at Tesla, you will participate in the full lifecycle of in-house data algorithms products and projects, from incubation to deployment, and from prototype to maturity. You will lead the entire process of data collection, cleaning, preprocessing, model training, and production deployment. The ideal candidate should be passionate about artificial intelligence and 3D vision technology, and stay abreast of the latest developments in the field.

This position primarily focuses on factory-related 3D machine vision applications, including automated visual quality inspection, robot guidance, vision-based dimension measurement (such as precise size verification, tolerance checking, and 3D shape analysis), object pose estimation, as well as point cloud processing and real-time perception in industrial automation scenarios.

Responsibilities
Machine Collaboration Control: Responsible for camera calibration, hand-eye calibration, and point cloud data processing (such as filtering, segmentation, detection, registration, and 6D pose estimation), collaborating with technologies like robots/PLC to drive the integration of 3D vision applications with production. Proactively conduct robot control and guidance to enhance production manufacturing efficiency.
3D Data Processing: Responsible for the collection, organization, filtering, and cleaning of large volumes of 3D data (such as point clouds and depth images) in machine vision projects. Must be proficient in handling vision-based dimension measurement tasks, including using laser triangulation or stereo vision methods for object size extraction, shape modeling, and accuracy calibration. Must be skilled in using tools like Python, C++, OpenCV, PCL (Point Cloud Library), Numpy, and Blender for 3D data processing.
Model Development: Responsible for data preprocessing, training, iteration, retraining of object detection, segmentation, and pose estimation models in 3D vision projects, as well as improving model accuracy and search tasks. In the field of vision-based dimension measurement, must develop and optimize related models (such as PointNet-based dimension estimation networks) to ensure measurement precision meets industrial standards (e.g., micron-level). Must have experience with frameworks like Python, C++, TensorFlow/PyTorch, and understand the application of common neural networks (such as CNN, PointNet, Transformer variants) in 3D vision. Familiar with tools like Pandas, MongoDB (Aggregation), Redis, and Kafka for model deployment.
Innovation: Stay sensitive to the latest 3D vision technologies and trends (such as real-time SLAM, Neural Radiance Fields NeRF, multimodal fusion), and propose innovative solutions to address industrial production challenges, for example, point cloud registration optimization in robot path planning or real-time tolerance detection optimization in vision-based dimension measurement to improve quality control efficiency.

Required 
Hold an advanced degree in computer science, mathematics, statistics, or a related discipline.
Strong development background in C++ and Python, capable of handling high-performance 3D vision algorithm implementations.
Strong background in 3D vision algorithms, including point cloud filtering, segmentation, detection, registration, and 6D pose estimation, as well as vision-based dimension measurement algorithms (such as 3D reconstruction and metrology techniques).
Experience in model development with frameworks like TensorFlow or PyTorch, and understanding of the principles and optimization of common neural networks (such as convolutional networks and point cloud networks) in 3D applications.
Data science capabilities, such as experience using tools like Pandas, Numpy, Matplotlib, and MongoDB Aggregation.
Knowledge of robot-related control technologies (such as ROS framework and path planning), and experience in factory 3D machine vision, including the application of vision-based dimension measurement in manufacturing quality control.

Preferred 
Practical project experience applying 3D vision technology in industrial, manufacturing, or robotics fields, such as automated guidance systems, defect detection, or vision-based dimension measurement systems.
Published research papers in relevant academic journals or conferences (such as CVPR, ICCV, IROS).
Experience with robot/PLC control, industrial cameras/laser sensors/lighting solutions.
Experience working in agile environments.
Experience in algorithm development competitions, such as ACM, Kaggle, or Robotics challenges.
Familiarity with multimodal data fusion (such as RGB-D combined with point clouds) or real-time 3D reconstruction technologies, particularly in vision-based dimension measurement applications (such as fusing sensor data to improve measurement accuracy).

Join Tesla, and you will have the opportunity to work in a dynamic and innovative environment, collaborating with world-class engineers and scientists to advance industrial automation and intelligent manufacturing through machine vision technology. If you are passionate about machine learning, artificial intelligence, and machine vision, and eager to realize your value in this cutting-edge field, we welcome you to join our team.

工作职责

数据算法团队在特斯拉工业智能研发方面扮演关键角色。我们通过自主搭建数据算法平台，赋能生产制造、供应链、销售、服务和充电网络等领域，将信息转化为高价值的数据资产，从而创造更优质的产品并提供完美的用户体验。

作为特斯拉应用软件团队的数据算法工程师，您将参与自研数据算法产品和项目的全生命周期，从孵化到落地，从雏形到成熟。您将领导数据的收集、清理、预处理、模型训练以及生产部署的全流程。理想候选人应对人工智能和3D视觉技术充满热情，并紧跟该领域的最新进展。

本职位主要聚焦于工厂相关的3D机器视觉应用，包括自动视觉质检、机器人引导、视觉尺寸测量（如精确尺寸验证、公差检查和3D形状分析）、物体姿态估计以及工业自动化场景中的点云处理和实时感知。

岗位职责
机器协同控制相关：负责相机标定、手眼标定、点云数据处理（如滤波、分割、检测、配准和6D位姿估计），配合机器人/PLC等技术，驱动3D视觉应用与生产协同。主动开展机器人控制和引导，促进生产制造效率。
3D数据处理相关：负责机器视觉项目中大批量3D数据（如点云、深度图像）的收集、整理、过滤和清洗。需熟练处理视觉尺寸测量任务，包括使用激光三角测量或立体视觉方法进行物体尺寸提取、形状建模和精度校准。需熟练使用Python、C++、OpenCV、PCL、Numpy、Blender等工具处理3D数据。
模型开发相关：负责3D视觉项目的物体检测、分割、姿态估计模型的数据预处理、训练、迭代、重训练，以及模型准确率提升和搜索任务。在视觉尺寸测量领域，需开发和优化相关模型（如基于PointNet的尺寸估计网络），确保测量精度达到工业标准（如微米级）。需具备Python、C++、TensorFlow/PyTorch等框架经验，并理解常用神经网络（如CNN、PointNet、Transformer变体）在3D视觉中的应用。熟悉Pandas、MongoDB（Aggregation）、Redis、Kafka等工具用于模型部署。
创新相关：对最新的3D视觉技术和趋势（如实时SLAM、神经辐射场NeRF、多模态融合）保持敏感，能够提出创新解决方案应对工业生产挑战，例如机器人路径规划中的点云配准优化或视觉尺寸测量中的实时公差检测优化，以提升质量控制效率。