具身智能数据闭环专家 - 火山引擎
Core
Design and iterate the data closed-loop system for embodied intelligence, managing the full lifecycle of multimodal data to support large model training and capability evolution.
Role type
Senior IC data platform engineer (embodied intelligence)
Builds
PB-scale multimodal data processing platforms and automated annotation toolchains for embodied intelligence models.
Domain
Embodied intelligence / Autonomous driving / Cloud AI
Deliverable
production ML models
Required skills
Distributed computing frameworks (Spark/Flink/Ray), Data lake technologies (Iceberg/Hudi), Python, Database design (MySQL/MongoDB/ES/Redis), Cloud-native architecture, Containerization, Multimodal data processing (video/point cloud/IMU), Active learning strategies, Synthetic data generation, Data quality governance.
Preferred skills
End-to-end or VLA large model data system experience, Experience building data closed-loops from 0 to 1 in autonomous driving.
Responsibilities
Architect and implement multimodal data platforms for storage, parsing, and retrieval; Design active learning strategies for high-value sample mining; Build automated annotation toolchains using large models; Establish data quality governance and version management; Generate synthetic and counterfactual data to expand scenario coverage.
Seniority
Senior, hands-on IC