Notes - Designing Machine Learning Systems by Chip Huyen

An Iterative Process for Production-Ready Applications.

Image credit: Unsplash

Overview of Machine Learning Systems

  • ML has found its way into almost every aspect of our lives: how we access information, how we communicate, how we work, how we find love etc.
  • May people when they hear Machine Learning System, think of just ML algorithms being used such as logistic regression or different types of neural networks. How ever algorithm is only a small part of an ML system in production.
  • The ML System includes:
    • Bussiness requirements.
    • The interface where users and developers interact with the system.
    • The data stack.
    • The logic for developing, monitoring and updating the model.
    • The infrastructure that enables the delivery of an ML system.
  • The relationship between MLOps and ML System Design
    • Ops comes from DevOps - Developments and Operations.
    • To operationalze something means to bring it into production, which includes deploying, monitoring, and maintaining it.
    • MLOps is a set of tools and best practices for bringing ML into production.
    • ML system design takes a system approach to MLOps which means that it considers and ML system holistically to ensure that all components and their stakeholders can work together to satisfy the specified objectives and requirements.
Viral Thakar
Viral Thakar
Machine Learning Engineer

My research interests include machine learning, computer vision and social innovations.