Keys to Building Machine Learning Systems

Discover essential engineering techniques for building successful machine learning systems in this 27-minute conference talk from GOTO Chicago 2020. Learn about the challenges organizations face in leveraging machine learning and how to overcome them through end-to-end engineering practices. Explore methods for collaborating with data scientists, common tool sets, and automation strategies. Understand the importance of short, iterative release cycles in data science projects and how to demonstrate business value early on. Gain insights on unifying data science and engineering teams, implementing software best practices, identifying model success criteria, solving simple problems first, automating processes from day one, and maintaining code quality through ongoing refactoring with tests. By the end of this talk, acquire the key knowledge needed to develop effective machine learning systems and drive AI adoption in enterprise environments.

Intro
What is machine learning?
AI adoption in the enterprise 2020
The fundamental problem in most organizations
Data science vs. traditional software
ML systems development
Unify data science & engineering teams
Use software best practices
Identify model success criteria metrics
Solve a simple problem first
Automate end-to-end on day 1
Ongoing refactoring with tests
Keys to building ML systems summary