Annotation Tech Jobs in California (NOW HIRING)

Job Summary:
XPENG is a leading smart technology company at the forefront of innovation, integrating advanced AI and autonomous driving technologies into its vehicles. They are seeking a Senior Staff Physical AI Data Algorithm Engineer to define and optimize the architecture of vehicle-cloud integrated data systems and lead the development of AI technologies for enhanced data processing and model performance.
Responsibilities:
• Responsible for the design and optimization of the vehicle-cloud integrated data closed-loop architecture: Build and maintain the full-link large closed-loop system from on-vehicle data upload to cloud training and simulation evaluation, ensuring efficient and secure data flow between the vehicle and the cloud to support rapid model iteration.
• Build and maintain the data closed-loop toolchain: Lead the selection, development and integration of modules such as data processing links, data mining, collection and annotation tools, and visualization tools to improve the automation level and processing efficiency of data from original collection to usable data sets.
• Establish data lineage and version management mechanisms: Design and implement a data lineage tracking system to achieve full-process traceability of data from production, processing to use; establish strict corresponding relationships between data sets, annotation versions, and model versions to support problem attribution and iterative backtracking.
• Explore the next-generation AI Agent-centric data closed-loop technology: Research and introduce AI Agent-based automated data processing and mining methods, explore the application of Agents in scenarios such as scene recognition, annotation assistance, and simulation use case generation, and promote the evolution of data closed-loop towards a higher level of intelligence.
• Support data work throughout the entire model development cycle: Deeply participate in the entire process of the model from data preparation, pre-training, fine-tuning, evaluation to on-board deployment and continuous optimization, understand the specific data needs of the model at each stage, and provide targeted data strategy support.
• Define high-quality data standards and guide data production: According to the key needs of different models at different stages (such as basic capability building, shortcoming repair, generalization improvement, etc.), clarify the characteristics of high-quality data (diversity, representativeness, scarcity, authenticity, etc.), guide data collection, cleaning and annotation work, and ensure model training effects.
Qualifications:
Required:
• Master's degree or above in Computer Science, Artificial Intelligence, Automation, Vehicle Engineering or related majors, with more than 3 years of work experience in multi-modal physical AI or AI data platform.
• In-depth understanding of the architecture and process of multi-modal physical AI data closed-loop, with integrated practical experience in on-vehicle data upload, cloud data processing, training and simulation integration.
• Familiar with the construction and use of data closed-loop toolchains, including data processing, mining, annotation, visualization and other modules, with relevant development or in-depth use experience.
• Have practical experience in the implementation of data lineage and version management, understand the importance of the association between data sets and model versions, and have a sense of data asset management.
• Have research or practical interest in the direction of AI Agent-centric data closed-loop, and have the ability to explore cutting-edge technologies.
• Familiar with the entire life cycle of model development, and deeply understand the key role of data in model performance (generalization, robustness, security).
• Able to analyze the data needs of the model at different stages, have the ability to define and evaluate high-quality data, and candidates with experience in guiding data production and annotation are preferred.
• Have good cross-team collaboration ability, able to work efficiently with algorithms, engineering, annotation, testing and other teams to promote the landing of data closed-loop.
Preferred:
• Candidates with experience in large-scale AI training data governance are preferred, including the construction of data standard systems, data quality governance, data asset management, cost and efficiency optimization, as well as practical experience in the implementation of massive multi-modal data production and circulation systems.
• Candidates with experience in guiding data production and annotation are preferred.
Company:
XPENG is a leading Chinese Smart EV company that designs, develops, manufactures, and markets Smart EVs that appeal to the large and growing base of technology-savvy middle-class consumers. Founded in 2014, the company is headquartered in Guangzhou, CHN, with a team of 10001+ employees. The company is currently Late Stage.