MLOps design

Concetti di MLOps

Folkert Stijnman

ML Engineer

Machine learning design

Added value

• Estimate the expected value
• ML is experimental and uncertain
• Aids in resource allocation, prioritization, and setting expectations

Business requirements

• End user
  • Speed
  • Accuracy
  • Transparency

Business requirements

• End user
  • Speed
  • Accuracy
  • Transparency
• Compliance and regulations

Business requirements

• End user
  • Speed
  • Accuracy
  • Transparency
• Compliance and regulations
• Budget
• Team size

Key metrics

Key metrics

Key metrics

Let's practice!

Concetti di MLOps