具身智能数据挖掘工程师（Seed） - Seed Model
Core
Design and implement multimodal data mining strategies and algorithms for robots to identify high-value samples, hard cases, and long-tail scenarios from massive datasets.
Role type
Senior IC data mining engineer (robotics)
Builds
Data value scoring systems, active learning loops, and model evaluation frameworks for robot training
Domain
Robotics + AI/ML
Deliverable
production ML models
Required skills
Python or Golang, multimodal data processing (image/video), statistical analysis, machine learning, distributed computing frameworks
Preferred skills
automatic annotation, data quality assurance, data diversity analysis, VLM/VLA training, active learning, hard case mining
Technologies
Ray, Daft, data lakes, data scheduling systems
Responsibilities
Design multimodal data mining strategies; Analyze data distribution and quality; Build active learning and data filtering loops; Participate in model evaluation体系建设 to identify failure patterns