Sr. Data Algorithm Engineer, CV

光学调试与视觉系统集成相关：负责工业相机选型、光源设计与调试、镜头选择、曝光/增益/触发等光学参数优化，以及2D相机标定。配合机器人/PLC等技术，驱动2D视觉应用与生产协同。 2D数据处理相关：负责机器视觉项目中大批量2D图像数据的收集、整理、标注、过滤和清洗。需熟练处理缺陷检测任务，包括图像预处理、数据增强、难例挖掘等。需熟练使用Python、C++、OpenCV、Numpy等工具处理2D图像数据。 模型开发相关：负责2D视觉项目的目标检测、语义/实例分割、缺陷分类/定位模型的数据预处理、训练、迭代、重训练，以及模型准确率、召回率和泛化能力提升。重点开发与优化少样本（few-shot）缺陷检测、异常检测（anomaly detection）相关模型，适应工业场景中缺陷样本稀缺的特点。需具备Python、C++、TensorFlow/PyTorch等框架经验，并理解常用神经网络（如CNN、Transformer及其变体）在2D视觉中的应用。 规模化部署与创新相关：利用AI Agent技术驱动视觉应用的规模化落地，包括自动化数据管道构建、模型自适应调优、跨产线快速部署、效果监控与持续迭代。对最新的2D视觉技术（few-shot learning、 foundation models、anomaly detection）和AI Agent趋势保持敏感，能够提出创新解决方案应对工业生产挑战，以提升质量控制效率和部署规模

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2D视觉算法实现。 具备很强的2D视觉算法背景，包括缺陷检测、分割、分类，以及少样本学习和异常检测算法。熟悉工业光学成像原理。 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、Transformer）在视觉中的原理和优化。 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。 了解工业相机、光源、镜头等硬件解决方案，具备工业制造的机器视觉经验，尤其是制造质量控制中的缺陷检测应用。具备较强的光学调试能力。 优先考虑 有实际将视觉技术应用于工业、制造领域的项目经验，特别是2D缺陷检测系统，few-shot缺陷检测项目经验为佳。 在相关领域的学术期刊或会议（如CVPR、ICCV）上发表过研究论文。 具有工业相机/光源/镜头解决方案的丰富经验，以及实际光学调试（lighting setup、相机参数优化）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等。 熟悉AI Agent、大模型在工业视觉规模化部署中的应用，或具备跨多产线/多场景快速落地视觉方案的经验。 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 industrial 2D machine vision technology, and stay abreast of the latest developments in the field. This position primarily focuses on factory-related 2D machine vision applications, including automated visual quality inspection (defect detection) and vision-based guidance. Strong emphasis is placed on 2D defect detection capabilities in industrial scenarios, few-shot defect detection project experience, and leveraging AI Agents to drive large-scale deployment of vision applications. Responsibilities Optical Tuning and Vision System Integration: Responsible for industrial camera selection, lighting design and debugging, lens selection, optimization of optical parameters (exposure, gain, trigger, etc.), and 2Dcamera calibration. Collaborate with robots/PLC technologies to integrate 2D vision applications with production. Proactively conduct vision-based guidance to enhance manufacturing efficiency. 2D Data Processing: Responsible for the collection, organization, annotation, filtering, and cleaning of large volumes of 2D image data in machine vision projects. Must be proficient in handling defect detection tasks, including image preprocessing, data augmentation, and hard example mining. Must be skilled in using tools like Python, C++, OpenCV, and Numpy for 2D image processing. Model Development: Responsible for data preprocessing, training, iteration, and retraining of object detection, semantic/instance segmentation, and defect classification/localization models in 2D vision projects, as well as improving model accuracy, recall, and generalization. Focus on developing and optimizing few-shot defect detection and anomaly detection models to address the scarcity of defect samples in industrial settings. Must have experience with frameworks like Python, C++,TensorFlow/PyTorch, and understand the application of common neural networks (such as CNNs and Transformer variants) in 2D vision. Familiar with tools like Pandas, MongoDB (Aggregation), Redis, and Kafka for model deployment. Large-Scale Deployment and Innovation: Leverage AI Agent technologies to drive large-scale rollout of vision applications, including automated data pipeline construction, adaptive model tuning, rapid cross-line deployment, monitoring, and continuous iteration. Stay sensitive to the latest 2D vision technologies (few-shot learning, foundation models, anomaly detection) and AI Agent trends, and propose innovative solutions to industrial production challenges to improve quality control efficiency and deployment scale. Required Qualifications 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 2D vision algorithm implementations. Strong background in 2D vision algorithms, including defect detection, segmentation, classification, few-shot learning, and anomaly detection. Familiarity with industrial optical imaging principles. 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 Transformers)in 2D applications. Data science capabilities, such as experience using tools like Pandas, Numpy, Matplotlib, and MongoDB Aggregation. Knowledge of industrial cameras, lighting, and lens solutions, with experience in factory 2D machine vision, especially defect detection in manufacturing quality control. Strong optical debugging and tuning skills required. Preferred Qualifications Practical project experience applying 2D vision technology in industrial or manufacturing fields, particularly2D defect detection systems. Few-shot defect detection project experience is highly preferred. Published research papers in relevant academic journals or conferences (such as CVPR, ICCV). Rich experience with industrial cameras, lighting, and lens solutions, as well as hands-on optical debugging(lighting setup, camera parameter optimization). Experience working in agile environments. Experience in algorithm development competitions, such as ACM or Kaggle. Familiarity with AI Agents and large models in the large-scale deployment of industrial vision applications, or experience in rapidly rolling out vision solutions across multiple production lines or scenarios.