Books and Guides 📗
Machine Learning Yearning
A concise, highly practical book on getting ML to work in practice.
As Andrew Ng himself says:
This book is focused not on teaching you ML algorithms, but on how to make ML algorithms work.
Rules of ML: Best Practices for ML Engineering
Drawing upon years of experience at Google, in this guide, Martin Zinkevich gives 43 “rules” related to different phases of Machine Learning projects.
Most of it is guided by the main thesis: “Most of the problems you will face (while doing ML in practice) are, in fact, engineering problems”.
To make great products: do machine learning like the great engineer you are, not like the great machine learning expert you aren’t.
ML Projects Guide
In this guide, Jeremy Jordan goes through all the phases of a typical machine learning project touching upon topics like collecting data, setting up the codebase, and “canarying”.
The goal of this document is to provide a common framework for approaching machine learning projects that can be referenced by practitioners.
Machine Learning: The High-Interest Credit Card of Technical Debt
Introducing ML to a system can give massive payoffs. But there is a price to pay as well in the form of technical debt. That is what this paper from Google addresses:
The goal of this paper is (to) highlight several machine learning specific risk factors and design patterns to be avoided or refactored where possible. These include boundary erosion, entanglement, hidden feedback loops, undeclared consumers, data dependencies, changes in the external world, and a variety of system-level anti-patterns.
Machine Learning in Practice: Guide
In this guide, “Seven Steps to Success: Machine Learning in Practice”, Daoud Clarke presents tips on getting ML to work in practice, including (interesting) sections on possible conversations with stakeholders.
Articles and Tutorials 📓
The step-by-step tutorial on building ML Products
This is a step-by-step guide to becoming an effective PM in an organization that leverages machine learning to achieve business goals.
Targetted toward project management, this series of tutorials touch on the non-technical side of things.
While ML is an incredibly technical space, many of the fundamentals you need to understand to maximize business impact have little to do with developing complex algorithms. They’re about ensuring you ask the right questions, understand the process of developing ML models, and structure an organization that fosters constant collaboration between disciplines rather than treating data science (the organization creating those models) as a “black box” that will magically generate insights.
ML in Real Life
Following tutorials and building some systems are well and good. But how do we get that to production, to the real world?
In this two-part series, Jade Abbott talks about some practical questions we encounter in practice.
The AI Hierarchy of Needs
Building AI capabilities in an organization isn’t a straightforward task. First, we need to assess the AI maturity. In this article, Monica Rogati talks about exactly that: how and why to determine a company’s readiness for AI.
How to Deliver on Machine Learning Projects
Making and maintaining progress in ML projects can be a challenge. In this article, the authors go over what they call “The ML Engineering Loop” consisting of four steps: i) Analyze ii) Select an approach iii) Implement iv) Measure.
The goal is “to cope with uncertainty and deliver great products quickly”.
How to apply ML to Business Problems
Coming from a more business-oriented angle, this article provides actionable insights for how to approach using ML to solve business problems.
A manifesto for Agile Data Science
Agile has been successful in “traditional” software development. But how does it look like for AI / Data Science applications? This article goes over how to think about agile for Data Science.
Practical advice for analysis of large, complex data sets
Data analysis is always an important part of any AI workflow. In this blog post from The Unofficial Google Data Science blog, Patrick Riley gives some advice to go about doing just that.
What do Machine Learning Practitioners actually do?
In thinking about what kind of processes to set up and directions to teach/learn for practical machine learning, it’s always good to know and think about what the practitioners actually do. In this blog post by Rachel Thomas, she goes over just that.
Getting Better at Machine Learning
How do we move beyond `model.fit(x, y)` to learn other things that are important in practice? In this article, Robert Chang shares some of his learnings from actually working on ML projects.
Podcasts 🎧
AI In Industry
Dan Faggella interviews data scientists, AI practitioners and AI leaders from companies around the world to figure out the applications and implications of AI and how it can be used to solve problems.
TWiML & AI
Each episode on “This Week in Machine Learning & AI” features an interview with an AI expert on a variety of topics.
The O’Reilly Data Show
The O’Reilly Data Show covers a host of ideas, techniques, and opportunities around big data, data science, and AI.
And that’s a wrap! If there are other resources I’ve missed out on (and I’m sure there are plenty), please comment below.
