OpenNash Jobs
Chery Automobile · 美国, 加利福尼亚, 旧金山 +3 more

Director, Data & AI Evaluation

美国, 加利福尼亚, 旧金山 +3 moreDirectorData & Analytics

Listed on Chery Automobile’s official careers site. Posted June 30, 2026. Check that it is still open and apply there.

Apply on Chery Automobile’s careers site →

Job description

"

  • Build a company-wide ""data factory"" — an end-to-end loop for data collection, processing, annotation, and feedback across edge and cloud
  • Design a full-stack automated evaluation system spanning foundation models, domain models, and application layers
  • Define data quality standards and model evaluation benchmarks; publish regular capability leaderboards
  • Develop intelligent data mining and augmentation techniques to maximize data utility in sparse scenarios
  • Lead AI compliance and ethical review programs to ensure data and model governance" "
  • Deep experience in large-scale data engineering and AI evaluation system design; fluency in data closed-loop processes
  • Proficiency in automated annotation, data cleaning, hard negative mining, and synthetic data generation
  • Led the design of an automated model evaluation platform supporting multi-dimensional assessment
  • Familiar with benchmark construction methodologies; prior experience building proprietary eval sets is a plus
  • Systematic understanding of data compliance, privacy protection, and AI ethics"

More jobs at Chery Automobile

Scientist, Auto-Labeling & Data Ops

美国, 加利福尼亚, 旧金山 +3 more

Staff Engineer, AI Data Engine

美国, 加利福尼亚, 旧金山 +3 more

Chief Scientist, Foundation Models

美国, 加利福尼亚, 旧金山 +3 more

Chief Scientist, World Models

美国, 加利福尼亚, 旧金山 +3 more

Director, AI Platform & Ecosystem

美国, 加利福尼亚, 旧金山 +3 more

Similar jobs at other companies

Testing & Evaluation Associate

JPMorganChase · Ciudad Autónoma de Buenos Aires, Argentina

Data Consultant Director

NTT · London, United Kingdom

OpenNash lists jobs from employers’ own careers sites and does not handle applications. Dates come from the employer where stated; “first seen” and “last checked” are our own observations.