Skip to main content

Introduction

Instruction iconInstruction and application
Complete iconIn Progress

Introduction​

Building a machine learning (ML) model is only the beginning. The real challenge is making sure it can grow with your data, users and business needs.

Scalability is about ensuring every stage of the ML pipeline adapts to growth without losing speed, accuracy or cost-efficiency. In this unit, you'll explore how to design ML systems that scale smoothly, operate efficiently and remain reliable under varying loads.

Finger pointing illustration
Lightbulb icon

Why does this unit matter?​

As ML applications move from experiments to production, they often face unexpected challenges when data volume increases or user demand spikes. A system that works for 1,000 users might collapse at 100,000.

Understanding scalability and capacity management allows you to anticipate these bottlenecks and proactively design solutions that maintain performance while controlling costs. This is essential for building sustainable and impactful ML systems.

Learning objectives​

By the end of this unit, you will be able to:

  • Analyse the impact of data characteristics on ML system scalability and resource requirements.
  • Apply capacity planning frameworks to estimate infrastructure needs for ML workloads.
  • Implement resource calculation strategies to balance performance, cost and reliability.
  • Develop supply chain risk mitigation plans for critical ML compute resources.
  • Evaluate auto-scaling strategies for various ML deployment scenarios.
Pause icon

Action item: Pause and think​

Reflect on a system you use daily. How do you think it handles a sudden 10x increase in users? What parts of the ML pipeline might be the first to struggle?

Questions & reflections
What scalability challenges have you observed in projects you've worked on?

Type your reflection here...